arXiv Science⌕ Search

arXiv · 2610.19655

Time-Inconsistent Singular Dividend Control with Regime Switching and Bankruptcy

Abstract

This paper studies infinite-horizon equilibrium dividend distribution for a firm whose surplus follows a diffusion under observable Markov regime switching and is absorbed at bankruptcy. Dividend payments are modeled as singular controls, and non-exponential discounting leads to time inconsistency. We formulate equilibrium control laws and establish a verification theorem for a general finite-state model with surplus-dependent coefficients. For a two-state model with constant positive drifts and volatilities and bi-exponential discounting, we prove the existence of a regime-dependent equilibrium barrier policy for arbitrary positive switching intensities. The policy retains surplus below the prevailing regime's threshold and distributes any excess immediately. Numerical experiments examine how shareholder patience and economic conditions affect dividend decisions. Over the parameter ranges considered, greater discount-rate dispersion at a fixed initial average rate raises both barriers, while switching speed can generate nonmonotone responses and reverse their ordering.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Xiaochen Chen, Guohui Guan, Zongxia Liang, Xiaodong Luo. 2026-10-14. Time-Inconsistent Singular Dividend Control with Regime Switching and Bankruptcy. https://arxiv.org/abs/2610.19655

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

Stochastic internal habit formation and optimality

Growth models with internal habit formation have been studied in various settings under the assumption of deterministic dynamics. The purpose of this paper is to explore a stochastic version of the model in Carroll et al. [1997, 2000], one the most influential on the subject. The goal is twofold: on one hand, to determine how far we can advance in the technical study of the model; on the other, to assess whether at least some of the deterministic outcomes remain valid in the stochastic setting. The resulting optimal control problem proves to be challenging, primarily due to the lack of concavity in the objective function. This feature is present in the model even in the deterministic case (see, e.g., Bambi and Gozzi [2020]). We develop an approach based on Dynamic Programming to establish several useful results, including the regularity of the solution to the corresponding HJB equation and a verification theorem. There results lay the groundwork for studying the model optimal paths and comparing them with the deterministic case.

math.OC↗

Optimal Transport on Lie Group Orbits

In its most general form, the optimal problem is an infinite-dimensional optimization problem, yet certain notable instances admit closed-form solutions. We identify the common source of this tractability as \textit{symmetry} and formalize it using Lie group theory. Fixing a Lie group action and a reference distribution, we study optimal transport between measures in the same orbit. For such pairs of measures, restricting the Monge problem to transport maps induced by the group action yields the Lie group orbit transport problem, an optimization over the stabilizer of the reference law that provides an upper bound on the Monge problem. Under suitable regularity conditions, a feasible group-induced map satisfying a $c$-convex first-order certificate attains this bound, solves the original Monge problem, and induces an optimal Kantorovich plan. For the quadratic cost, we establish a structural criterion based on a global Cartan decomposition. Exploiting the nonuniqueness of orbit representations, the criterion constructs an optimal map through a path of group transformations with self-adjoint derivatives. Once verified for an action and reference law, it applies to every pair of orbit measures with finite second moments. For affine-induced actions, Cartan structure on the linear image suffices, and the means and covariance operators determine the optimal map and value explicitly. The framework recovers classical elliptical transport formulas and yields closed-form solutions on congruence orbits generated by absolutely continuous, spectrally invariant reference laws, including Wishart, inverse-Wishart, and matrix beta type~II families with fixed shape parameters. It also recovers the common-copula optimal coupling for multivariate normal sinh-arcsinh laws with unit tailweight parameters sharing a Gaussian copula, and evaluates the optimal transport value in closed form.

math.OC↗

Robust Parallel Machine Scheduling: A Comparison of Robustness Measures and Local Search Approaches

In many real-world scheduling problems, the processing times of tasks are subject to uncertainty. This makes it essential to design schedules that are robust against disruptions. We distinguish between two types of robustness: quality-robust schedules keep a stable objective value under disruptions, while solution-robust schedules keep a stable schedule structure. We study how to generate schedules for the Stochastic Parallel Machine Scheduling Problem (SPMSP) with precedence constraints and release dates that are both quality- and solution-robust. To this end, we compare several robustness measures and local search algorithms that use these measures. We start with an extensive simulation study where we determine the correlation of 13 existing and 6 new robustness measures with both quality and solution robustness. Next, we use the measures with high robustness correlation as the objective in our local search algorithm to find the best way to generate schedules containing both job-machine assignments as well as buffer assignments. We use a new dynamic objective weight within our local search algorithms, which actively changes the objective weights in each iteration. We found that the best way to generate robust schedules is to use robustness measures that model the processing times with normal distribution approximations and to combine job placement and buffer assignment simultaneously.

math.OC↗