Marginal Posterior Simulation via Higher-order Tail Area Approximations