An iterated Azéma-Yor type embedding for finitely many marginals (Q2406561): Difference between revisions

The authors consider an \(n\)-marginal Skorokhod embedding problem and construct an explicit solution with desirable optimal properties. To describe the problem, consider a standard Brownian motion \(B\) and a sequence of probability measures \(\mu_1,\dots,\mu_n\). A solution is a sequence of stopping times \(\tau_1\leq \cdots\leq \tau_n \) such that \(B_{\tau_i}\sim \mu_i\), \(1\leq i\leq n\), and \(B_{t\wedge \tau_n}\), \(t\geq 0\), is a uniformly integrable martingale. A solution may exist only if all \(\mu_i\) are centered and their sequence is in convex order. The problem is solved under additional technical assumptions. The nuances of the solution of the problem and its complexity in comparison with the case \(n=1\) are clearly explained. Applications to robust pricing and hedging of lookback options are discussed.