A barrier option of American type (Q1596351)

This paper deals with the stock price-per-share \(S(t)\) which is assumed to satisfy the standard model \[ dS(t)=S(t)[r\,dt+\sigma \,dW_{0}(t)], S(0)=x\in (0,h), \] of Merton (1973) and Black and Scholes (1973), with \(r>0\) the prevalent interest rate of the risk-free asset (bank account), \(\sigma>0\) the volatility of the stock, and \(W_{0}(t)\) a Brownian motion under the risk-neutral equivalent martingale measure. The pricing and hedging problems for the ''up-and-out'' Barrier Put-Option of American type, with payoff \[ Y(t)=(q-S(t))^{+}{\mathbf 1}_{\{t<\tau_{h}\}}, 0\leq \infty, \] are solved in closed form, both with and without constraints on the short-selling of the stock. Here \(h>0\) is the barrier and \(q\in (0,h)\) is the strike-price of the option, whereas \(\tau_{h}:=\inf\{t\geq 0/S(t)>h\}\) is the time when the option becomes ``knocked-out''. The constrained case leads to a stochastic optimization problem of mixed optimal stopping/singular control type. This is reduced to a variational inequality which is then solved explicitly in two qualitatively separate cases, according to a certain compatibility condition among the market coefficients and the constraint. The relevant counterparts of singular stochastic control for the pricing of European-type barrier options was first brought out in \textit{U. Wystup} [Doctoral Dissertation, Carnegie Mellon University (1997)]. An analysis of ``down-and-out'' barrier call-options, of both European and American type, is carried out in \textit{R. C. Merton} [Bell. J. Econ. Manage. Sci. 4, 141--183 (1973)].