The Financial Bubble Model with L\'evy Jump Processes

We present a mathematical extension of the Berestycki--Monneau--Scheinkman (BMS) model for speculative financial bubbles, by incorporating jump-diffusion dynamics. While the original BMS framework assumes that investor disagreement evolves along strictly continuous paths, our model captures sudden, discontinuous shifts in market sentiment through an independent L\'evy jump process. Under this generalized heterogeneous beliefs framework, we show via optimal stopping theory and the It\^o--L\'evy formula that the speculative bubble premium satisfies a non-local partial integro-differential equation (PIDE) with a moving obstacle. The first main contribution of this paper is the development of a complete viscosity solution theory for this specific non-local obstacle problem. We establish a comparison principle using a doubling-of-variables argument adapted for non-local jump integrals, and we rigorously prove the existence and uniqueness of the bubble price via Perron's method by constructing explicit, continuous sub- and supersolution barriers. Furthermore, we design a strictly monotone Implicit-Explicit (IMEX) finite difference scheme to compute the bubble premium. By extending the Barles-Souganidis framework, we prove that our discrete matrix operator preserves the M-matrix property and that the numerical scheme converges locally uniformly to the unique viscosity solution under a spatial-dependent Courant--Friedrichs--Lewy (CFL) condition. Finally, we implement this scheme via a PSOR-Picard algorithm, and simulate the bubble across four finite- and infinite-activity L\'evy models.