Paper: arXiv 2502.11052

Authors: Yike Wang, Yusha Chen, Jingzhen Liu

Abstract

This paper is devoted to time-consistent control problems of portfolio selection with strictly monotone mean-variance preferences. These preferences are variational modifications of the conventional mean-variance preferences, and remain time-inconsistent as in mean-variance optimization problems. To tackle the time-inconsistency, we study the Nash equilibrium controls of both the open-loop type and the closed-loop type, and characterize them within a random parameter setting. The problem is reduced to solving a flow of forward-backward stochastic differential equations for open-loop equilibria, and to solving extended Hamilton-Jacobi-Bellman equations for closed-loop equilibria. In particular, we derive semi-closed-form solutions for these two types of equilibria under a deterministic parameter setting. Both solutions are represented by the same function, which is independent of wealth state and random path. This function can be expressed as the conventional time-consistent mean-variance portfolio strategy multiplied by a factor greater than one. Furthermore, we find that the state-independent closed-loop Nash equilibrium control is a strong equilibrium strategy in a constant parameter setting only when the interest rate is sufficiently large.

Complexity vs Empirical Score

  • Math Complexity: 9.0/10
  • Empirical Rigor: 2.0/10
  • Quadrant: Lab Rats — high math complexity, limited empirical testing

Why this score: The paper relies heavily on advanced stochastic control theory, forward-backward SDEs, and extended HJB equations, indicating high mathematical complexity. However, the analysis is purely theoretical with no empirical data, backtesting, or implementation details, resulting in low empirical rigor.

Research Flowchart

  flowchart TD
  A["Research Goal<br>Find time-consistent<br>portfolio selection<br>for strict monotone<br>mean-variance preference"] --> B{"Methodology<br>Nash Equilibrium Analysis"}

  B --> C["Type 1: Open-Loop<br>Solve Forward-Backward<br>Stochastic Differential Equations"]
  B --> D["Type 2: Closed-Loop<br>Solve Extended<br>HJB Equations"]

  C --> E["Compute Equilibria"]
  D --> E

  E --> F["Key Findings/Outcomes<br>1. Semi-closed solutions<br>2. Strategy = Conventional MV<br>Strategy × Factor > 1<br>3. Wealth-Independent<br>4. Strong equilibrium<br>requires high interest rate"]

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 Nash equilibrium controls (open-loop and closed-loop) for monotone mean-variance (MMV) portfolio selection and derives semi‑closed‑form solutions under deterministic parameter settings.

Source sentence
To address this time-inconsistency, we consider Nash equilibrium controls of both open-loop and closed-loop types, and characterize them within a random parameter setting. In particular, we derive semi-closed-form solutions for both types of equilibria under a deterministic parameter setting, and both solutions share the same representation, which is independent of the wealth state and the random path.

Assumes 6

  • Model parameters (interest rate r, volatility σ, risk premium ϑ, and monotonicity parameter ζ) are deterministic. model
    Source sentence
    Assume that (r, σ, ϑ, ζ) are deterministic. Then, ¯π is an ONEC if ¯πt = βtϑt θσt e− R T t rvdv, P −a.s., a.e. t ∈[0, T], (23) where the continuous and decreasing function β : [0, T] →[1, +∞) is the unique solution of the following integral equation with βT = 1: 1 = βtΦ ◦Ψ  1 −ζ ( R T t |βsϑs|2ds) 1 2  , t ∈[0, T).
  • All parameters are constant over time in the numerical illustration. model
    Source sentence
    Assuming the parameters (r, σ, ϑ, ζ) are constant, we perform a calibration exercise using monthly S&P 500 stock index from January 1990 to December 2025.
  • The distribution of the terminal wealth deviation is uniquely determined by its moments. statistical
    Source sentence
    A key underlying assumption is that the distribution of Xt,x,Π T −Et[Xt,x,Π T ] is uniquely determined by its moments.
  • Continuity and strict monotonicity of the function R_λ ensure existence and uniqueness of λ. model
    Source sentence
    Notably, the existence and uniqueness of λXπ T ,θ,ζ arise from the continuity and strict monotonicity of R λ −∞P(Xπ T + ζ θ ≤s)ds on λ ∈(ess inf{Xπ T + ζ θ}, +∞).
Show 2 more
  • The terminal value function M(T, x, λ) is twice differentiable for t∈[0,T). model
    Source sentence
    See Section A.6. □ Given that the terminal value M(T, x, λ) = ζ + θ(λ −x −ζ θ)+ is not continuously differentiable, we have indeed assumed that M(t, ·, λ) is twice differentiable for t ∈[0, T), analogous to deriving the Black-Scholes-Merton equation for option pricing (e.g., (Shreve, 2004, Section 4.5)).
  • The wealth process follows the standard stochastic differential equation with constant drift ϑ and diffusion σ. market
    Source sentence
    Assume that (r, σ, ϑ, ζ) are deterministic.

Data

  • Monthly S&P 500 stock index · WIND database market-prices commercial
    January 1990 to December 2025 · S&P 500 constituents
    Source sentence
    Assuming the parameters (r, σ, ϑ, ζ) are constant, we perform a calibration exercise using monthly S&P 500 stock index from January 1990 to December 2025. Data are sourced from the WIND database and supplemented by the 30-year U.S. Treasury bond annual yields for the corresponding period, sourced from the U.S. Department of the Treasury (U.S. Department of the Treasury, 2026).
  • 30‑year U.S. Treasury bond annual yields · U.S. Department of the Treasury macro public
    January 1990 to December 2025
    Source sentence
    Data are sourced from the WIND database and supplemented by the 30-year U.S. Treasury bond annual yields for the corresponding period, sourced from the U.S. Department of the Treasury (U.S. Department of the Treasury, 2026).

Recognised sources: Wind

Costs

Unclear No cost treatment was found in the passages the extractor read; that is not evidence that costs are not modeled.

Finds 5

  • The investment amount under the MMV equilibrium exceeds that under the MV equilibrium, and the gap narrows over time. performance
    Source sentence
    We show that the investment amount under the MMV equilibrium exceeds that under the MV equilibrium, and the gap narrows over time.
  • The closed-loop Nash equilibrium is a strong equilibrium only when the interest rate is sufficiently large, whereas the open-loop Nash equilibrium is always a strong equilibrium. theoretical
    Source sentence
    Furthermore, under a constant parameter setting, we find that the derived closed-loop Nash equilibrium control is a strong equilibrium strategy only when the interest rate is sufficiently large, whereas the derived open-loop Nash equilibrium control is necessarily a strong equilibrium strategy.
  • Both open-loop and closed-loop equilibria have the same semi‑closed‑form representation, independent of wealth level and random path. theoretical
    Source sentence
    In particular, we derive semi-closed-form solutions for both types of equilibria under a deterministic parameter setting, and both solutions share the same representation, which is independent of the wealth state and the random path. The two equilibrium strategies are found to be identical, which are independent of the wealth level and random path.
  • Equilibrium investment in the risky asset under SMMV preference is higher than under MV preference because β_t > 1. performance
    Source sentence
    As noted in Section 5, since βt > 1, the equilibrium investment in the risky asset with the SMMV preference ¯πt is higher than that under the MV preference ¯πMV t .
  • Semi‑closed‑form solutions for the deterministic‑parameter case are obtained (Theorem 5.1 and 5.3). theoretical figure from elsewhere in the paper
    Source sentence
    See Section A.7. □ 5 Semi-closed-form solution for deterministic parameters In the case that the model parameters are all deterministic, we obtain a semi-closed-form expression of the ONEC, which is a deterministic function of t. Assume that (r, σ, ϑ, ζ) are deterministic.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 28 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 (52)

  1. Dynamic asset allocation under mean-reverting returns, stochastic interest rates, and inflation uncertainty: A re popular recommendations consistent with rational behavior? (2004) — link
  2. Interest Rate Statistics: D aily Treasury Par Yield Curve Rates (2026) — link
  3. Characterizations of closed-loop equilibrium solutions for dynamic mean-variance optimization problems (2017) — link
  4. Dynamic Mean-Variance Asset Allocation (2010) — link
  5. Convex Analysis and Monotone Operator Theory in H ilbert Spaces (2017) — link
  6. L$^P$ solutions of backward stochastic differential equations (2003) — link
  7. Dynamic asset allocation under inflation (2002) — link
  8. On time-inconsistent stochastic control in continuous time (2017) — link
and 44 more
  1. Mean-variance portfolio optimization with state dependent risk aversion (2014) — link
  2. Semimartingale theory of monotone mean-variance portfolio allocation (2020) — link
  3. On the computation of optimal monotone mean–variance portfolios via truncated quadratic utility (2012) — link
  4. Dynamically optimal portfolios for monotone mean–variance preferences — → 2503.08272
  5. Stochastic Calculus and Applications (2015) — link
  6. Mean-variance theory in complete markets (1982) — link
  7. The Mathematics of Arbitrage (2006) — link
  8. Controlled M arkov processes and viscosity solutions (2006) — link
  9. Consistent Plans (1980) — link
  10. On the Equilibrium Strategies for Time-Inconsistent Problems in Continuous Time (2021) — link
  11. Time-inconsistent Stochastic Linear-Quadratic Control (2012) — link
  12. Time-inconsistent Stochastic Linear-Quadratic Control: Characterization And Uniqueness Of Equilibrium (2017) — link
  13. Constrained monotone mean-variance problem with random coefficients — → 2212.14188
  14. Viscosity solutions of fully nonlinear second-order elliptic partial differential equations (1990) — link
  15. Is Mean-Variance Analysis Vacuous: Or was Beta Still Born? (1997) — link
  16. Mathematical methods for financial market (2009) — link
  17. Some extensions of Ito’s formula (1981) — link
  18. Optimal reinsurance and investment strategies for an insurer under monotone mean-variance criterion (2021) — link
  19. Optimal investment and reinsurance policies for the C ram 'e r- L undberg risk model under monotone mean-variance preference (2023) — link
  20. Continuous-time monotone mean-variance portfolio selection (2022) — link
  21. Comparison between mean-variance and monotone mean-variance preferences under jump diffusion and stochastic factor model (2024) — link
  22. Comparison between mean-variance and monotone mean-variance preferences in general markets: A new perspective (2025) — link
  23. Optimal dynamic portfolio selection: Multi-period mean-variance formulation (2000) — link
  24. Dynamic Portfolio Selection for Nonlinear Law-Dependent Preferences — → 2311.06745
  25. Portfolio Selection (1952) — link
  26. Ambiguity aversion, robustness, and the variational representation of preferences (2006) — link
  27. Portfolio selection with monotone mean-variance preferences (2009) — link
  28. Envelope Theorems for Arbitrary Choice Sets (2002) — link
  29. Convex Functions, Monotone Operators and Differentiability (1993) — link
  30. Consistent Planning (1968) — link
  31. On the Existence of a Consistent Course of Action when Tastes are Changing (1973) — link
  32. Stochastic calculus for finance II : C ontinuous-time model (2004) — link
  33. Myopia and Inconsistency in Dynamic Utility Maximization (1955) — link
  34. A note on monotone mean-variance preferences for continuous processes (2020) — link
  35. Optimal strategies and values for monotone and classical mean-variance preferences coincide when asset prices are continuous (2024) — link
  36. Constrained monotone mean-variance investment-reinsurance under the C ram'er-Lundberg model with random coefficients (2024) — link
  37. Cone-constrained Monotone Mean-Variance Portfolio Selection Under Diffusion Models (2022) — link
  38. Continuous-Time portfolio choice under monotone mean-variance preferences–stochastic factor case (2019) — link
  39. The stochastic Fubini theorem revisited (2012) — link
  40. Time-inconsistent optimal control problems and the equilibrium HJB equation (2012) — link
  41. Stochastic Controls: Hamiltonian Systems and HJB equations (1999) — link
  42. Continuous-Time Mean-Variance Portfolio Selection: A Stochastic LQ Framework (2000) — link
  43. On stochastic control problems with higher-order moments — → 2412.13521
  44. Strictly monotone mean-variance preferences with applications to portfolio selection — → 2412.13523

5 of 52 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.