Title:Time-changed Markov processes and coupled non-local equations

Abstract:Motivated by a financial valuation problem on an asset-pricing model with dependent trade duration and returns, in this paper we study coupled fully non-local equations, where a linear non-local operator jointly acts on the time and space variables. We prove existence and uniqueness of the solution. Existence is established by providing a stochastic representation based on anomalous processes constructed as a time change via the undershooting of an independent subordinator. This leads to general non-stepped processes with intervals of constancy representing a sticky or trapping effect (i.e., constant price in financial applications). Our theory allows these intervals to be dependent on the immediately subsequent jump. A maximum principle is then proved and used to derive uniqueness. Based on these general results, we consider a particular case: a non-local analog of the Black and Scholes equation, addressing the problem of determining the seasoned price of a derivative security.