pith. sign in

arxiv: 2605.24833 · v2 · pith:G4WWRHGAnew · submitted 2026-05-24 · 🧮 math.OC · q-fin.RM

Controlled McKean--Vlasov Contagion with State-Dependent Killing

Aoxin Zhang , Yingzhe Wang This is my paper

Pith reviewed 2026-06-30 00:14 UTC · model grok-4.3

classification 🧮 math.OC q-fin.RM
keywords McKean-Vlasov equationscomparison principlesHamilton-Jacobi-Bellman equationsstate-dependent killingcontagion modelsmean-field limitsparticle approximationspropagation of chaos
0
0 comments X

The pith

Comparison principle holds for the two-population killed-particle HJB on decomposed alive-measure and cemetery state spaces.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops controlled McKean-Vlasov models that incorporate contagion, common noise, loss feedback, and state-dependent killing across interacting populations. Its central result is a comparison principle for the associated two-population Hamilton-Jacobi-Bellman equation defined on a state space split between living sub-probability measures and cemetery masses. The argument proceeds by combining a Wasserstein smooth-gauge comparison with a killing-jump absorption estimate that accounts for mass transfer into the cemetery. The work further derives a multi-population mean-field limit, an explicit first-order convergence rate for particles, conditional propagation of chaos, and controlled well-posedness, plus a steep-killing link to absorbing-boundary defaults. These results matter for rigorous analysis of absorption and default in large-scale controlled interacting systems.

Core claim

The central claim is that a comparison principle holds for the two-population killed-particle HJB on a decomposed state space of alive sub-probability measures and cemetery masses. The proof combines a Wasserstein smooth-gauge comparison argument with a killing-jump absorption estimate for mass transfer into the cemetery state. The paper also establishes a multi-population mean-field limit, an explicit first-order particle convergence rate, conditional propagation of chaos, controlled well-posedness, and a steep-killing bridge to absorbing-boundary default, with supporting finite-particle tests and a two-population HJB feedback experiment.

What carries the argument

The decomposed state space of alive sub-probability measures and cemetery masses together with the killing-jump absorption estimate that preserves comparison under mass transfer.

If this is right

  • A multi-population mean-field limit exists for the controlled system.
  • An explicit first-order particle convergence rate holds.
  • Conditional propagation of chaos is valid.
  • Controlled well-posedness follows for the interacting particle system.
  • A steep-killing regime connects the model to absorbing-boundary default.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The comparison principle could support viscosity-solution existence results for related controlled jump equations.
  • Similar absorption estimates might extend the method to models with additional jump types or non-constant killing intensities.
  • The decomposed-space technique may inform numerical schemes that track cemetery mass separately in high-dimensional control problems.
  • The mean-field and convergence results could apply to contagion settings in epidemiology or network failure models with removal states.

Load-bearing premise

The state space can be decomposed into alive sub-probability measures and cemetery masses such that the killing mechanism produces a well-defined jump absorption estimate preserving the comparison property.

What would settle it

An explicit counter-example or numerical test in which the comparison principle fails for a concrete killing rate that produces mass transfer violating the absorption estimate.

Figures

Figures reproduced from arXiv: 2605.24833 by Aoxin Zhang, Yingzhe Wang.

Figure 1
Figure 1. Figure 1: Finite-type matrix scaling and risk–cost frontier in the near-critical region. [PITH_FULL_IMAGE:figures/full_fig_p017_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Particle-number convergence validation for Theorem 2. [PITH_FULL_IMAGE:figures/full_fig_p018_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Adjacency sparsity patterns and weighted in-exposure distributions for the three sparse [PITH_FULL_IMAGE:figures/full_fig_p019_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Representative subgraph layouts for the three sparse network classes. Node positions are [PITH_FULL_IMAGE:figures/full_fig_p019_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Sparse graph systems versus aggregate matrix approximations under exposure scaling. [PITH_FULL_IMAGE:figures/full_fig_p020_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Exposure multipliers at which different network structures reach the graph cascade [PITH_FULL_IMAGE:figures/full_fig_p020_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Confidence bands for cascade probabilities and tail gaps with 500 replications in the [PITH_FULL_IMAGE:figures/full_fig_p021_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Theorem 3 diagnostics for the matrix-approximation boundary. [PITH_FULL_IMAGE:figures/full_fig_p022_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Common-noise effects on sparse graph risk and graph-pressure control. [PITH_FULL_IMAGE:figures/full_fig_p023_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Control strategies and cascade probabilities in representative sparse graph regions. [PITH_FULL_IMAGE:figures/full_fig_p023_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Tail risk–control cost relation in representative sparse graph regions. [PITH_FULL_IMAGE:figures/full_fig_p024_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Cascade probability comparison between type-uniform control and graph-pressure [PITH_FULL_IMAGE:figures/full_fig_p025_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: K = 2 HJB multi-grid and Richardson-type diagnostics. The Richardson-type panel in [PITH_FULL_IMAGE:figures/full_fig_p026_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Asymmetric loss-coordinate refinement diagnostic for the [PITH_FULL_IMAGE:figures/full_fig_p027_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: (x1, x2) heat maps of K = 2 HJB feedback control on the fixed-loss slice. 27 [PITH_FULL_IMAGE:figures/full_fig_p027_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: shows control slices at a fixed loss state. Control intensity is more concentrated in high-pressure regions, and core-type control increases earlier as system pressure rises. This structure is consistent with the sign interpretation of the projected fixed-point marginal control condition stated in the appendix HJB characterization: when the local pressure of a type has a larger marginal contribution to fu… view at source ↗
Figure 17
Figure 17. Figure 17: further gives time slices of the HJB feedback. Early control is closer to a preventive allocation. As the terminal time approaches, feedback becomes increasingly concentrated near the boundary and in regions where the loss state has already deteriorated. This dynamic form is consistent with the marginal benefit term in the Hamiltonian: control has higher value during periods in which future losses can sti… view at source ↗
Figure 18
Figure 18. Figure 18: Population loss paths in the K = 2 core–periphery system. In this explicit core–periphery case, no control and the graph-pressure threshold rule both have mean loss 0.361, 95% loss 0.377, cascade probability 1.000, and zero cost because the two-type projection does not expose within-type pressure differences. Type-uniform control reduces mean loss to 0.200, 95% loss to 0.210, and cascade probability to 0.… view at source ↗
Figure 19
Figure 19. Figure 19: Risk–cost frontiers under the parameter grid for graph-pressure threshold control. [PITH_FULL_IMAGE:figures/full_fig_p030_19.png] view at source ↗
Figure 20
Figure 20. Figure 20: Sparse graph robustness under longer horizons and time-varying networks. [PITH_FULL_IMAGE:figures/full_fig_p031_20.png] view at source ↗
Figure 21
Figure 21. Figure 21: Effects of the killing-function slope γ and the baseline kill intensity λbase on terminal mean loss and 95% tail loss. Cascade probability is saturated on this grid and is summarized in the text. 32 [PITH_FULL_IMAGE:figures/full_fig_p032_21.png] view at source ↗
Figure 22
Figure 22. Figure 22: Robustness under additional network families and larger sparse systems. [PITH_FULL_IMAGE:figures/full_fig_p033_22.png] view at source ↗
Figure 23
Figure 23. Figure 23: Tail loss and cascade probability in K = 3 and K = 4 finite-type matrix systems. 14 External-data calibrated network We combine the preceding synthetic network mechanism with public regulatory data to construct a semi-realistic calibrated core–periphery network. The calibration inputs come from the EBA 2024 EU-wide Transparency Exercise and the official disclosure packages of the EBA Pillar 3 Data Hub / E… view at source ↗
Figure 24
Figure 24. Figure 24: Jurisdictional distribution and asset-size proxies in the EBA and Pillar 3 calibrated [PITH_FULL_IMAGE:figures/full_fig_p036_24.png] view at source ↗
Figure 25
Figure 25. Figure 25: Risk and control results in the EBA and Pillar 3 calibrated synthetic core–periphery [PITH_FULL_IMAGE:figures/full_fig_p037_25.png] view at source ↗
Figure 26
Figure 26. Figure 26: Average loss paths in the EBA and Pillar 3 calibrated synthetic core–periphery network. [PITH_FULL_IMAGE:figures/full_fig_p037_26.png] view at source ↗
Figure 27
Figure 27. Figure 27: Multi-instance calibrated synthetic networks under EBA and Pillar 3 constraints. [PITH_FULL_IMAGE:figures/full_fig_p038_27.png] view at source ↗
Figure 28
Figure 28. Figure 28: Candidate points in the K = 2 HJB numerical calibration grid. Point color represents the kill-intensity multiplier, and point size represents the nonzero-control share. SM2 Proofs for the mean-field limit, quantitative rate, and propa￾gation of chaos Lemma SM2.1 (Conditional empirical LLN for killed diffusion paths). Fix a type k. Given the common-noise path W0 and a common-noise adapted type-level contro… view at source ↗
read the original abstract

We study controlled McKean--Vlasov contagion with state-dependent killing, common noise, loss feedback, and interacting populations. The main result is a comparison principle for the two-population killed-particle HJB on a decomposed state space of alive sub-probability measures and cemetery masses. The proof combines a Wasserstein smooth-gauge comparison argument with a killing-jump absorption estimate for mass transfer into the cemetery state. We also establish a multi-population mean-field limit, an explicit first-order particle convergence rate, conditional propagation of chaos, controlled well-posedness, and a steep-killing bridge to absorbing-boundary default. Finite-particle convergence tests and a two-population HJB feedback experiment illustrate the theory.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript develops a theory of controlled McKean-Vlasov contagion processes with state-dependent killing, common noise, loss feedback, and interacting populations. Its central claim is a comparison principle for the two-population killed-particle HJB equation posed on a decomposed state space consisting of alive sub-probability measures and cemetery masses; the proof combines a Wasserstein smooth-gauge comparison argument with a killing-jump absorption estimate. Additional results include a multi-population mean-field limit, an explicit first-order particle convergence rate, conditional propagation of chaos, controlled well-posedness, and a steep-killing bridge to absorbing-boundary default, illustrated by finite-particle convergence tests and a two-population HJB feedback experiment.

Significance. If the comparison principle holds under the stated assumptions, the work would supply a useful analytic tool for uniqueness and viscosity-solution approaches in mean-field control problems that incorporate killing and contagion. The explicit convergence rates, the propagation-of-chaos result, and the numerical experiments constitute concrete strengths that go beyond purely existential statements.

major comments (2)
  1. [Abstract / main result paragraph] Abstract / main result paragraph: the comparison principle is stated for the decomposed state space of alive sub-probability measures and cemetery masses, yet the manuscript supplies neither the precise functional form of the HJB operator nor the assumptions (e.g., regularity or growth conditions) imposed on the state-dependent killing rate that are required for the killing-jump absorption estimate to preserve the comparison property.
  2. [State-space decomposition (main result paragraph)] State-space decomposition (main result paragraph): the claim that the killing mechanism produces a well-defined jump absorption estimate that preserves comparison rests on the decomposition being compatible with the Wasserstein smooth-gauge metric; the text does not verify that the cemetery-mass coordinate does not introduce discontinuities or violate the gauge properties used in the comparison argument.
minor comments (2)
  1. [Numerical experiments] The numerical section mentions finite-particle convergence tests and a two-population HJB feedback experiment but does not report the discretization scheme, the number of particles, or the specific parameter values employed.
  2. [Notation] Notation for the cemetery mass and the alive sub-probability measures should be introduced once and used consistently throughout the statements of the mean-field limit and the convergence-rate theorems.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive suggestions. The comments correctly identify points where additional explicit statements would strengthen the presentation of the main result. We address each major comment below and will revise the manuscript accordingly.

read point-by-point responses

  1. Referee: [Abstract / main result paragraph] Abstract / main result paragraph: the comparison principle is stated for the decomposed state space of alive sub-probability measures and cemetery masses, yet the manuscript supplies neither the precise functional form of the HJB operator nor the assumptions (e.g., regularity or growth conditions) imposed on the state-dependent killing rate that are required for the killing-jump absorption estimate to preserve the comparison property.

    Authors: We agree that the abstract and main-result paragraph would benefit from an explicit statement of the HJB operator and the precise regularity/growth conditions on the killing rate. These are given in Sections 2--3 (Lipschitz continuity in the measure variable with linear growth, uniform boundedness, and measurability with respect to the common noise), which guarantee the killing-jump absorption estimate. In the revision we will insert a concise description of the operator and the key assumptions into the abstract and the statement of the comparison principle. revision: yes

  2. Referee: [State-space decomposition (main result paragraph)] State-space decomposition (main result paragraph): the claim that the killing mechanism produces a well-defined jump absorption estimate that preserves comparison rests on the decomposition being compatible with the Wasserstein smooth-gauge metric; the text does not verify that the cemetery-mass coordinate does not introduce discontinuities or violate the gauge properties used in the comparison argument.

    Authors: The decomposition treats the cemetery mass as a deterministic, absolutely continuous coordinate driven by the integral of the killing rate; the Wasserstein smooth-gauge is extended by a Lipschitz term in this coordinate. The proof in Section 4 already uses this extension, but an explicit verification that the gauge inequalities and continuity properties are preserved is not isolated as a separate lemma. We will add a short remark or auxiliary lemma confirming that the cemetery coordinate introduces neither discontinuities nor violations of the gauge properties. revision: yes

Circularity Check

0 steps flagged

No significant circularity

full rationale

The paper states a comparison principle for the killed-particle HJB as its main result, proved by combining a Wasserstein smooth-gauge comparison argument with a killing-jump absorption estimate on the decomposed state space. No equations, parameters, or derivations are shown that reduce by construction to fitted inputs, self-definitions, or load-bearing self-citations. The result is presented as an independent theorem with supporting elements (mean-field limit, convergence rates) that do not exhibit the enumerated circularity patterns. The derivation chain is self-contained against external mathematical benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only the abstract is available; no explicit free parameters, axioms, or invented entities can be extracted or verified.

pith-pipeline@v0.9.1-grok · 5647 in / 1084 out tokens · 30278 ms · 2026-06-30T00:14:33.608637+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

62 extracted references · 45 canonical work pages

  1. [1]

    Eisenberg and T

    L. Eisenberg and T. H. Noe,Systemic risk in financial systems, Manage. Sci., 47 (2001), pp. 236–249,https://doi.org/10.1287/mnsc.47.2.236.9835

    work page doi:10.1287/mnsc.47.2.236.9835 2001

  2. [2]

    Allen and D

    F. Allen and D. Gale,Financial contagion, J. Polit. Econ., 108 (2000), pp. 1–33,https: //doi.org/10.1086/262109

    work page doi:10.1086/262109 2000

  3. [3]

    Gai and S

    P. Gai and S. Kapadia,Contagion in financial networks, Proc. R. Soc. A, 466 (2010), pp. 2401–2423,https://doi.org/10.1098/rspa.2009.0410

    work page doi:10.1098/rspa.2009.0410 2010

  4. [4]

    Acemoglu, A

    D. Acemoglu, A. Ozdaglar, and A. Tahbaz-Salehi,Systemic risk and stability in financial networks, Amer. Econ. Rev., 105 (2015), pp. 564–608,https://doi.org/10.1257/ aer.20130456

    2015

  5. [5]

    Elliott, B

    M. Elliott, B. Golub, and M. O. Jackson,Financial networks and contagion, Amer. Econ. Rev., 104 (2014), pp. 3115–3153,https://doi.org/10.1257/aer.104.10.3115

    work page doi:10.1257/aer.104.10.3115 2014

  6. [6]

    Glasserman and H

    P. Glasserman and H. P. Young,Contagion in financial networks, J. Econ. Lit., 54 (2016), pp. 779–831,https://doi.org/10.1257/jel.20151228

    work page doi:10.1257/jel.20151228 2016

  7. [7]

    Hambly, S

    B. Hambly, S. Ledger, and A. Søjmark,A McKean–Vlasov equation with positive feedback and blow-ups, Ann. Appl. Probab., 29 (2019), pp. 2338–2373,https://doi.org/10.1214/ 18-AAP1455

    2019

  8. [8]

    Hambly and A

    B. Hambly and A. Søjmark,An SPDE model for systemic risk with endogenous contagion, Finance Stoch., 23 (2019), pp. 535–594,https://doi.org/10.1007/s00780-019-00396-1

    work page doi:10.1007/s00780-019-00396-1 2019

  9. [9]

    Hambly and P

    B. Hambly and P. Jettkant,Control of McKean–Vlasov SDEs with contagion through killing at a state-dependent intensity, Ann. Appl. Probab., 35 (2025), pp. 4381–4438,https: //doi.org/10.1214/25-AAP2226

    work page doi:10.1214/25-aap2226 2025

  10. [10]

    V. V. Acharya, L. H. Pedersen, T. Philippon, and M. Richardson,Measuring systemic risk, Rev. Financ. Stud., 30 (2017), pp. 2–47,https://doi.org/10.1093/rfs/hhw088

    work page doi:10.1093/rfs/hhw088 2017

  11. [11]

    Adrian and M

    T. Adrian and M. K. Brunnermeier,CoVaR, Amer. Econ. Rev., 106 (2016), pp. 1705–1741, https://doi.org/10.1257/aer.20120555

    work page doi:10.1257/aer.20120555 2016

  12. [12]

    L. C. G. Rogers and L. A. M. Veraart,Failure and rescue in an interbank network, Manage. Sci., 59 (2013), pp. 882–898,https://doi.org/10.1287/mnsc.1120.1569. 115

    work page doi:10.1287/mnsc.1120.1569 2013

  13. [13]

    Cifuentes, G

    R. Cifuentes, G. Ferrucci, and H. S. Shin,Liquidity risk and contagion, J. Eur. Econom. Assoc., 3 (2005), pp. 556–566,https://doi.org/10.1162/jeea.2005.3.2-3.556

    work page doi:10.1162/jeea.2005.3.2-3.556 2005

  14. [14]

    P. Gai, A. Haldane, and S. Kapadia,Complexity, concentration and contagion, J. Monetary Econ., 58 (2011), pp. 453–470,https://doi.org/10.1016/j.jmoneco.2011.05.005

    work page doi:10.1016/j.jmoneco.2011.05.005 2011

  15. [15]

    R. J. Caballero and A. Simsek,Fire sales in a model of complexity, J. Finance, 68 (2013), pp. 2549–2587,https://doi.org/10.1111/jofi.12087

    work page doi:10.1111/jofi.12087 2013

  16. [16]

    R. M. May and N. Arinaminpathy,Systemic risk: The dynamics of model banking systems, J. R. Soc. Interface, 7 (2010), pp. 823–838,https://doi.org/10.1098/rsif.2009.0359

    work page doi:10.1098/rsif.2009.0359 2010

  17. [17]

    M. Boss, H. Elsinger, M. Summer, and S. Thurner,Network topology of the interbank market, Quant. Finance, 4 (2004), pp. 677–684, https://doi.org/10.1080/ 14697680400020325

    2004

  18. [18]

    E. Nier, J. Yang, T. Yorulmazer, and A. Alentorn,Network models and financial stability, J. Econom. Dynam. Control, 31 (2007), pp. 2033–2060,https://doi.org/10.1016/ j.jedc.2007.01.014

    2007

  19. [19]

    R. Cont, A. Moussa, and E. B. Santos,Network structure and systemic risk in banking systems, in Handbook on Systemic Risk, J.-P. Fouque and J. A. Langsam, eds., Cambridge University Press, 2013, pp. 327–368,https://doi.org/10.1017/CBO9781139151184.018

    work page doi:10.1017/cbo9781139151184.018 2013

  20. [20]

    Battiston, D

    S. Battiston, D. Delli Gatti, M. Gallegati, B. Greenwald, and J. E. Stiglitz, Liaisons dangereuses: Increasing connectivity, risk sharing, and systemic risk, J. Econom. Dynam. Control, 36 (2012), pp. 1121–1141,https://doi.org/10.1016/j.jedc.2012.04.001

    work page doi:10.1016/j.jedc.2012.04.001 2012

  21. [21]

    Glasserman and H

    P. Glasserman and H. P. Young,How likely is contagion in financial networks?, J. Banking Finance, 50 (2015), pp. 383–399,https://doi.org/10.1016/j.jbankfin.2014.02.006

    work page doi:10.1016/j.jbankfin.2014.02.006 2015

  22. [22]

    Caccioli, J

    F. Caccioli, J. D. Farmer, N. Foti, and D. Rockmore,Overlapping portfolios, contagion, and financial stability, J. Econom. Dynam. Control, 51 (2015), pp. 50–63,https://doi.org/ 10.1016/j.jedc.2014.09.041

    work page doi:10.1016/j.jedc.2014.09.041 2015

  23. [23]

    Capponi, P.-C

    A. Capponi, P.-C. Chen, and D. D. Yao,Liability concentration and systemic losses in financial networks, Oper. Res., 64 (2016), pp. 1121–1134,https://doi.org/10.1287/opre. 2015.1402

    work page doi:10.1287/opre 2016

  24. [24]

    Gandy and L

    A. Gandy and L. A. M. Veraart,A bayesian methodology for systemic risk assessment in financial networks, Manage. Sci., 63 (2017), pp. 4428–4446,https://doi.org/10.1287/mnsc. 2016.2546

    work page doi:10.1287/mnsc 2017

  25. [25]

    Demirer, F

    M. Demirer, F. X. Diebold, L. Liu, and K. Yilmaz,Estimating global bank network connectedness, J. Appl. Econometrics, 33 (2018), pp. 1–15,https://doi.org/10.1002/jae. 2585

    work page doi:10.1002/jae 2018

  26. [26]

    Amini, R

    H. Amini, R. Cont, and A. Minca,Resilience to contagion in financial networks, Math. Finance, 26 (2016), pp. 329–365,https://doi.org/10.1111/mafi.12051

    work page doi:10.1111/mafi.12051 2016

  27. [27]

    J. P. Gleeson,High-accuracy approximation of binary-state dynamics on networks, Phys. Rev. Lett., 107 (2011), p. 068701,https://doi.org/10.1103/PhysRevLett.107.068701. 116

    work page doi:10.1103/physrevlett.107.068701 2011

  28. [28]

    T. R. Hurd, J. P. Gleeson, and S. Melnik,A framework for analyzing contagion in assortative banking networks, PLoS ONE, 12 (2017), p. e0170579,https://doi.org/10.1371/ journal.pone.0170579

    2017

  29. [29]

    Fouque and J

    J.-P. Fouque and J. A. Langsam, eds.,Handbook on Systemic Risk, Cambridge University Press, 2013,https://doi.org/10.1017/CBO9781139151184

    work page doi:10.1017/cbo9781139151184 2013

  30. [30]

    Fouque and T

    J.-P. Fouque and T. Ichiba,Stability in a model of interbank lending, SIAM J. Financial Math., 4 (2013), pp. 784–803,https://doi.org/10.1137/110841096

    work page doi:10.1137/110841096 2013

  31. [31]

    Garnier, G

    J. Garnier, G. Papanicolaou, and T.-W. Yang,Large deviations for a mean field model of systemic risk, SIAM J. Financial Math., 4 (2013), pp. 151–184,https://doi.org/10.1137/ 12087387X

    2013

  32. [32]

    Ledger and A

    S. Ledger and A. Søjmark,Uniqueness for contagious McKean–Vlasov systems in the weak feedback regime, Bull. London Math. Soc., 52 (2020), pp. 448–463,https://doi.org/10.1112/ blms.12337

    2020

  33. [33]

    Delarue, J

    F. Delarue, J. Inglis, S. Rubenthaler, and E. Tanré,Particle systems with a singular mean-field self-excitation. application to neuronal networks, Stochastic Process. Appl., 125 (2015), pp. 2451–2492,https://doi.org/10.1016/j.spa.2015.01.007

    work page doi:10.1016/j.spa.2015.01.007 2015

  34. [34]

    Sznitman,Topics in propagation of chaos, in École d’Été de Probabilités de Saint-Flour XIX—1989, vol

    A.-S. Sznitman,Topics in propagation of chaos, in École d’Été de Probabilités de Saint-Flour XIX—1989, vol. 1464 of Lecture Notes in Mathematics, Springer, Berlin, 1991, pp. 165–251, https://doi.org/10.1007/BFb0085169

    work page doi:10.1007/bfb0085169 1989

  35. [35]

    Budhiraja, P

    A. Budhiraja, P. Dupuis, and M. Fischer,Large deviation properties of weakly interacting processes via weak convergence methods, Ann. Probab., 40 (2012), pp. 74–102,https://doi. org/10.1214/10-AOP616

    work page doi:10.1214/10-aop616 2012

  36. [36]

    Feinstein and A

    Z. Feinstein and A. Søjmark,Dynamic default contagion in heterogeneous interbank systems, SIAM J. Financial Math., 12 (2021), pp. SC83–SC97,https://doi.org/10.1137/20M1376765

    work page doi:10.1137/20m1376765 2021

  37. [37]

    Feinstein and A

    Z. Feinstein and A. Søjmark,Contagious McKean–Vlasov systems with heterogeneous impact and exposure, Finance Stoch., 27 (2023), pp. 663–711,https://doi.org/10.1007/ s00780-023-00504-2

    2023

  38. [38]

    Huang, R

    M. Huang, R. P. Malhamé, and P. E. Caines,Large population stochastic dynamic games: Closed-loop McKean–Vlasov systems and the Nash certainty equivalence principle, Commun. Inf. Syst., 6 (2006), pp. 221–252,https://doi.org/10.4310/CIS.2006.v6.n3.a5

    work page doi:10.4310/cis.2006.v6.n3.a5 2006

  39. [39]

    R. A. Carmona, J.-P. Fouque, and L.-H. Sun,Mean field games and systemic risk, Commun. Math. Sci., 13 (2015), pp. 911–933,https://doi.org/10.4310/CMS.2015.v13.n4. a4

    work page doi:10.4310/cms.2015.v13.n4 2015

  40. [40]

    Carmona and F

    R. Carmona and F. Delarue,Probabilistic Theory of Mean Field Games with Applications I–II, Springer, 2018

    2018

  41. [41]

    Fabbri, F

    G. Fabbri, F. Gozzi, and A. Swiech,Stochastic Optimal Control in Infinite Dimension, Springer, 2017,https://doi.org/10.1007/978-3-319-53067-3

    work page doi:10.1007/978-3-319-53067-3 2017

  42. [42]

    M. G. Crandall, H. Ishii, and P.-L. Lions,User’s guide to viscosity solutions of second order partial differential equations, Bull. Amer. Math. Soc., 27 (1992), pp. 1–67,https: //doi.org/10.1090/S0273-0979-1992-00266-5. 117

    work page doi:10.1090/s0273-0979-1992-00266-5 1992

  43. [43]

    Cosso, F

    A. Cosso, F. Gozzi, I. Kharroubi, H. Pham, and M. Rosestolato,Master bellman equation in the Wasserstein space: Uniqueness of viscosity solutions, Trans. Amer. Math. Soc., 377 (2024), pp. 31–83,https://doi.org/10.1090/tran/8986

    work page doi:10.1090/tran/8986 2024

  44. [44]

    Gangbo, A

    W. Gangbo, A. R. Mészáros, C. Mou, and J. Zhang,Mean field games master equations with nonseparable hamiltonians and displacement monotonicity, Ann. Probab., 50 (2022), pp. 2178–2217,https://doi.org/10.1214/22-AOP1580

    work page doi:10.1214/22-aop1580 2022

  45. [45]

    Chassagneux, D

    J.-F. Chassagneux, D. Crisan, and F. Delarue,Numerical method for FBSDEs of McKean–Vlasov type, Ann. Appl. Probab., 29 (2019), pp. 1640–1684,https://doi.org/10. 1214/18-AAP1429

    2019

  46. [46]

    J. Han, A. Jentzen, and W. E,Solving high-dimensional partial differential equations using deep learning, Proc. Natl. Acad. Sci. USA, 115 (2018), pp. 8505–8510,https://doi.org/10. 1073/pnas.1718942115. [47]L. Lovasz,Large Networks and Graph Limits, American Mathematical Society, 2012

    2018

  47. [47]

    Benjamini and O

    I. Benjamini and O. Schramm,Recurrence of distributional limits of finite planar graphs, Electron. J. Probab., 6 (2001), pp. 1–13,https://doi.org/10.1214/EJP.v6-96

    work page doi:10.1214/ejp.v6-96 2001

  48. [48]

    Borgs, J

    C. Borgs, J. T. Chayes, H. Cohn, and N. Holden,Sparse exchangeable graphs and their limits via Graphon processes, J. Mach. Learn. Res., 18 (2018), pp. 1–71. JMLR article 210; available athttps://jmlr.org/papers/v18/16-421.html

    2018

  49. [49]

    Delattre, G

    S. Delattre, G. Giacomin, and E. Luçon,A note on dynamical models on random graphs and Fokker–Planck equations, J. Stat. Phys., 165 (2016), pp. 785–798,https://doi.org/10. 1007/s10955-016-1652-3

    2016

  50. [50]

    R. I. Oliveira and G. H. Reis,Interacting diffusions on random graphs with diverging average degrees: Hydrodynamics and large deviations, J. Stat. Phys., 176 (2019), pp. 1057–1087, https://doi.org/10.1007/s10955-019-02332-1

    work page doi:10.1007/s10955-019-02332-1 2019

  51. [51]

    R. I. Oliveira, G. H. Reis, and L. M. Stolerman,Interacting diffusions on sparse graphs: Hydrodynamics from local weak limits, Electron. J. Probab., 25 (2020), pp. 1–35, https://doi.org/10.1214/20-EJP505

    work page doi:10.1214/20-ejp505 2020

  52. [52]

    Parise and A

    F. Parise and A. Ozdaglar,Graphon games: A statistical framework for network games and interventions, Econometrica, 91 (2023), pp. 191–225,https://doi.org/10.3982/ECTA17564

    work page doi:10.3982/ecta17564 2023

  53. [53]

    P. E. Caines and M. Huang,Graphon mean field games and their equations, SIAM J. Control Optim., 59 (2021), pp. 4373–4399,https://doi.org/10.1137/20M136373X

    work page doi:10.1137/20m136373x 2021

  54. [54]

    Carmona, D

    R. Carmona, D. B. Cooney, C. V. Graves, and M. Laurière,Stochastic Graphon games: I. the static case, Math. Oper. Res., 47 (2022), pp. 750–778,https://doi.org/10. 1287/moor.2021.1148

    work page arXiv 2022

  55. [55]

    Aurell, R

    A. Aurell, R. Carmona, G. Dayanıklı, and M. Laurière,Finite state Graphon games with applications to epidemics, Dyn. Games Appl., 12 (2022), pp. 49–81,https://doi.org/10. 1007/s13235-021-00410-2

    2022

  56. [56]

    Bayraktar, S

    E. Bayraktar, S. Chakraborty, and R. Wu,Graphon mean field systems, Ann. Appl. Probab., 33 (2023), pp. 3587–3619,https://doi.org/10.1214/22-AAP1901. 118

    work page doi:10.1214/22-aap1901 2023

  57. [57]

    Lacker and A

    D. Lacker and A. Soret,A case study on stochastic games on large graphs in mean field and sparse regimes, Math. Oper. Res., 47 (2022), pp. 1530–1565,https://doi.org/10.1287/ moor.2021.1179

    work page arXiv 2022

  58. [58]

    R. J. Bendel,Sharp transitions and systemic risk in sparse financial networks, arXiv preprint arXiv:2601.04096, (2026),https://doi.org/10.48550/arXiv.2601.04096

    work page doi:10.48550/arxiv.2601.04096 2026

  59. [59]

    Pub- lic data release, https://www.eba.europa.eu/risk-and-data-analysis/risk-analysis/ eu-wide-transparency-exercise/2024-eu-wide-transparency-exercise

    European Banking Authority,2024 eu-wide transparency exercise, 2024. Pub- lic data release, https://www.eba.europa.eu/risk-and-data-analysis/risk-analysis/ eu-wide-transparency-exercise/2024-eu-wide-transparency-exercise

    2024

  60. [60]

    European Data Ac- cess Portal disclosure platform, https://www.eba.europa.eu/risk-and-data-analysis/ pillar-3-data-hub

    European Banking Authority,Pillar 3 data hub, 2026. European Data Ac- cess Portal disclosure platform, https://www.eba.europa.eu/risk-and-data-analysis/ pillar-3-data-hub

    2026

  61. [61]

    Public-domain map data,https: //www.naturalearthdata.com/

    Natural Earth,Natural earth vector map data, 2024. Public-domain map data,https: //www.naturalearthdata.com/

    2024

  62. [62]

    A. W. van der Vaart and J. A. Wellner,Weak Convergence and Empirical Processes: With Applications to Statistics, Springer Series in Statistics, Springer, New York, 1996,https: //doi.org/10.1007/978-1-4757-2545-2. 119

    work page doi:10.1007/978-1-4757-2545-2 1996