Paper: arXiv 2310.18658
Authors: Weihuan Huang
Abstract
CoVaR (conditional value-at-risk) is a crucial measure for assessing financial systemic risk, which is defined as a conditional quantile of a random variable, conditioned on other random variables reaching specific quantiles. It enables the measurement of risk associated with a particular node in financial networks, taking into account the simultaneous influence of risks from multiple correlated nodes. However, estimating CoVaR presents challenges due to the unobservability of the multivariate-quantiles condition. To address the challenges, we propose a two-step nonparametric estimation approach based on Monte-Carlo simulation data. In the first step, we estimate the unobservable multivariate-quantiles using order statistics. In the second step, we employ a kernel method to estimate the conditional quantile conditional on the order statistics. We establish the consistency and asymptotic normality of the two-step estimator, along with a bandwidth selection method. The results demonstrate that, under a mild restriction on the bandwidth, the estimation error arising from the first step can be ignored. Consequently, the asymptotic results depend solely on the estimation error of the second step, as if the multivariate-quantiles in the condition were observable. Numerical experiments demonstrate the favorable performance of the two-step estimator.
Complexity vs Empirical Score
- Math Complexity: 9.0/10
- Empirical Rigor: 3.0/10
- Quadrant: Lab Rats — high math complexity, limited empirical testing
Why this score: The paper presents advanced statistical theory including consistency proofs, asymptotic normality derivations, and complex kernel-based nonparametric methods, indicating high mathematical complexity. While it uses Monte Carlo simulations for validation, the lack of backtesting against real market data, absence of implementation code, and reliance on synthetic data limit its empirical readiness for trading.
Research Flowchart
flowchart TD Start["Research Goal<br>Estimate CoVaR for<br>Systemic Risk in Financial Networks"] --> Problem["Problem<br>Unobservable Multivariate-Quantiles<br>Condition for CoVaR"] Problem --> Method["Proposed Method<br>Two-Step Nonparametric Estimation"] Method --> Data["Input<br>Monte-Carlo Simulation Data"] Data --> Step1["Step 1: First-Step Estimation<br>Estimate Multivariate-Quantiles<br>Using Order Statistics"] Step1 --> Step2["Step 2: Kernel Estimation<br>Estimate Conditional Quantile<br>Conditioned on Order Statistics"] Step2 --> Results["Outcomes<br>Consistent & Asymptotically Normal Estimator<br>Favorable Performance in Numerical Experiments"] style Start fill:#e1f5fe,stroke:#01579b,stroke-width:2px style Problem fill:#fff3e0,stroke:#e65100,stroke-width:2px style Method fill:#e8f5e8,stroke:#2e7d32,stroke-width:2px style Data fill:#f3e5f5,stroke:#4a148c,stroke-width:2px style Step1 fill:#e8f5e8,stroke:#2e7d32,stroke-width:2px style Step2 fill:#e8f5e8,stroke:#2e7d32,stroke-width:2px style Results fill:#e1f5fe,stroke:#01579b,stroke-width:2px
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 a two-step nonparametric estimation method for CoVaR with a multivariate-quantiles condition, using order statistics in the first step and a kernel estimator in the second step.
Source sentence
To address the challenges, we propose a two-step nonparametric estimation approach based on Monte-Carlo simulation data. In the first step, we estimate the unobservable multivariate-quantiles using order statistics.
Assumes 3
- The random variables X1,…,Xm,Y are continuous and have positive, continuous densities; the joint density f(x,y) and marginal fX(x) are continuous and bounded on their support. statistical
Source sentence
Let qα = (q1,α1, . . . , qm,αm)⊤where qj,αj is the αj-quantile of Xj, j = 1, . . . , m. Assumption 1. Then, (i) both fX(x) and f(x, y) are positive and continuous in X and X × Y respectively; (ii) for any y ∈Y and ℓ∈Nm such that |ℓ| ≤1, we have DℓfX(x) and Dℓg(x, y) are continuous in X; (iii) for x ∈Rm and ℓ∈Nm such that |ℓ| ≤1, we have Dℓg(x, y) is uniformly continuous in y ∈Y; (iv) for y ∈Y and ℓ∈Nm such that |ℓ|
- The kernel function K is a bounded, symmetric, compactly supported probability density satisfying standard moment conditions. statistical
Source sentence
The kernel function K is a bounded, symmetric and compactly supported probability density function that satisfying (i) tK(t) →0 as |t| →∞; (ii) ´ R K(t)dt = 1; (iii) ´ R |K(t)|dt < ∞; (iv) ´ R tK(t)dt = 0; (v) ´ R t2K(t)dt < ∞; (vi) for any ℓ∈Nm such that |ℓ| ≤1, we have [DℓQm j=1 K(xj)]2 is bounded and integrable, and lim∥x∥→∞∥x∥· DℓQm j=1 K(xj) = 0.
- All m bandwidths approach zero as the sample size n → ∞, and nΔn → ∞ so that the bias diminishes and the convergence rate is (nΔn)^{-1/2}. statistical
Source sentence
This assumption requires that all m bandwidths approach zero as the sample size n →∞, ensuring that the bias of the kernel estimation diminishes as n →∞. Additionally, the assumption n∆n →∞imposes a condition that the band-widths cannot decrease too rapidly, guaranteeing the convergence of the kernel estimation with a rate of (n∆n)−1/2.
Data
- Monte‑Carlo simulation data synthetic unknown
Source sentence
To overcome the challenges, this paper proposes a two-step nonparametric method based on Monte-Carlo simulation. Suppose that we have observed an independent and identically distributed (i.i.d.) simulation sample from a validated model, denoted by n�X1,1, . . . , X1,m, Y1 , �X2,1, . . . , X2,m, Y2 , . . . , �Xn,1, . . . , Xn,m, Yn o .
Costs
Not applicable
Finds 5
- The two-step estimator is consistent and asymptotically normal, with the first-step error negligible under a mild bandwidth restriction. theoretical
Source sentence
We establish the consistency and asymptotic normality of the two-step estimator, along with a bandwidth selection method. The results demonstrate that, under a mild restriction on the bandwidth, the estimation error arising from the first step can be ignored.
- Numerical experiments show favorable performance of the two-step estimator. performance
Source sentence
Numerical experiments demonstrate the favorable performance of the two-step estimator.
- As the sample size increases, the bias, standard deviation, and RMSE of the two-step estimator approach zero. performance
Source sentence
The table demonstrates that as the sample size tends to infinity, the bias, SD, and RMSE of the two-step estimator approach zero.
- The two-step estimator outperforms the batching estimator, exhibiting smaller bias, SD, and RMSE. performance
Source sentence
Upon closer examination of Table 2, we observe that the two-step estimator outperforms the batching estimator, exhibiting smaller bias, SD, and RMSE.
- Coverage probability of 95% confidence intervals converges to the nominal level of 0.95 with increasing sample size. statistical
Source sentence
Moreover, the CP converges to the nominal level of 0.95 with increasing sample size.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 20 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.
Cited References (97)
- Approximation Theorems of Mathematical Statistics (1980)
- CoVaR (2016)
- Measuring Systemic Risk (2017)
- Estimating quantile sensitivities (2009)
- Kernel estimation of quantile sensitivities (2009)
- Systemic risk measurement: Multivariate GARCH estimation of CoVaR (2013)
- On dependence consistency of CoVaR and some other systemic risk measures (2014)
- Time-varying systemic risk: Evidence from a dynamic copula model of CDS spreads (2018)
and 89 more
- Bayesian inference for CoVaR (2013)
- Multivariate Markov-Switching models and tail risk interdependence (2018) — link
- Multi- CoVaR and shapley value: A systemic risk measure (2013)
- Measuring systemic risk in the European banking sector: A copula CoVaR approach (2018)
- Non-Gaussian models for CoVaR estimation (2022)
- An extreme value approach to CoVaR estimation (2022) — link
- A note on quantiles in large samples (1966)
- Simulating sensitivities of conditional value at risk (2009)
- Conditional Monte Carlo estimation of quantile sensitivities (2009)
- Simulating risk contributions of credit portfolios (2015)
- Monte Carlo Methods in Financial Engineering (2004)
- Importance sampling for Monte Carlo estimation of quantiles (1996)
- Variance reduction techniques for estimating value-at-risk (2000)
- Probability: Theory and Examples, 5th Edition (2019)
- Measure and probability (1995)
- Financial prediction with constrained tail risk (2007)
- Portfolio value-at-risk with heavy-tailed risk factors (2002)
- Risk Management and Financial Institutions, 5th Edition (2018)
- Stochastic calculus for finance II: Continuous-time models (2004)
- Continuous-time finance (1992)
- Probability-1 (2016)
- Quantile regression (2001)
- Nested simulation in portfolio risk measurement (2010)
- Monte Carlo simulation in financial engineering (2007)
- Asymptotic representations for importance-sampling estimators of value-at-risk and conditional value-at-risk (2010)
- Confidence intervals for quantiles when applying variance-reduction techniques (2012)
- Monte Carlo methods for value-at-risk and conditional value-at-risk: A review (2014)
- The Multivariate Normal Distribution (2012)
- An Introduction to Multivariate Statistical Analysis 2nd ed. (1984)
- An overview of value at risk (1997)
- Financial risk and heavy tails (2003)
- Portfolio credit risk with extremal dependence: Asymptotic analysis and efficient simulation (2008)
- An application of the GARCH-t model on Central European Stock Returns (2004)
- GARCH forecasting performance under different distribution assumptions (2006)
- Value at Risk: The New Benchmark for Managing Financial Risk, 3rd Edition (2000)
- Kernel smoothing for nested estimation with application to portfolio risk measurement (2017)
- Confidence intervals for quantiles using sectioning when applying variance-reduction techniques (2014)
- Simulation Modeling and Analysis, 5th Edition (2015)
- Asymptotic Statistics (2000)
- How likely is contagion in financial networks? (2015)
- A survey of systemic risk analytics (2012)
- Challenges in identifying and measuring systemic risk (2013)
- Systemic Risk, Crises, and Macroprudential Regulation (2015)
- Systemic risk in financial networks: A survey (2021)
- Systemic risk 10 years later (2018)
- Contagion in financial networks (2016)
- Stochastic Differential Equations: An Introduction with Applications (2003)
- Introduction to Probability Models, 12th Edition (2019)
- Learning-based robust optimization: Procedures and statistical guarantees (2021)
- Nonparametric Econometrics: Theory and Practice (2007)
- Monte- C arlo estimation of CoVaR (2022) — link
- Pricing interest-rate-derivative securities (1990)
- Time-dependent diffusion models for term structure dynamics (2003)
- The valuation of options for alternative stochastic processes (1976)
- Computing the constant elasticity of variance option pricing formula (1989)
- The constant elasticity of variance option pricing model (1996)
- The pricing of commodity contracts (1976)
- The constant elasticity of variance model and its implications for option pricing (1980)
- Dynamic consumption and portfolio choice with stochastic volatility in incomplete markets (2005)
- A closed-form solution for options with stochastic volatility with applications to bond and currency options (1993)
- Dynamic nonmyopic portfolio behavior (1996)
- Optimum consumption and portfolio rules in a continuous-time model (1975)
- Portfolio and consumption decisions under mean-reverting returns: An exact solution for complete markets (2002)
- Consumption and portfolio decisions when expected returns are time varying (1999)
- Nonparametric Econometrics (1999)
- Nonparametric and Semiparametric Models (2004)
- On estimation of a probability density function and mode (1962)
- Harmonic Analysis and The Theory of Probability (2005)
- Estimation of a probability density function and its derivatives (1967)
- Estimation of a probability density function and its derivatives (1969)
- Kernel and nearest-neighbor estimation of a conditional quantile (1990)
- A kernel estimator of a conditional quantile (1996)
- Nonparametric conditional quantile estimation: A locally weighted quantile kernel approach (2017)
- A generalized L1-approach for a kernel estimator of conditional quantile with functional regressors: Consistency and asymptotic normality (2009)
- A smooth conditional quantile estimator and related applications of conditional empirical processes (1991)
- Nonparametric estimation and inference on conditional quantile processes (2015)
- Bootstrap Methods and Their Application (1997)
- TENET : Tail-event driven network risk (2016)
- A Functional Coefficient VAR Model for Dynamic Quantiles With an Application for Constructing Financial Network (2020) — link
- Systemic risk in financial networks: The effects of asymptotic independence (2023) — link
- Quantile Regression (2005)
- Financial networks and contagion (2014)
- Networks in finance (2009)
- Systemic risk and stability in financial networks (2015)
- The formation of financial networks (2016)
- Financial network systemic risk contributions (2015)
- Systemic risk in financial systems (2001)
- VAR for VaR : Measuring tail dependence using multivariate regression quantiles (2015)
- Pointwise improvement of multivariate kernel density estimates (1998)
0 of 97 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.