US intervention: Assessing the probability of success

The martingale nature of exchange rate changes insures that intervention often will appear successful in terms of altering or moderating exchange rate movements, even if intervention were ineffective and undertaken randomly. I provide evidence that intervention generally lacks forecast value, except under a weak leaning-against-the-wind criterion. When I condition the probability of success by various aspects or techniques of intervention, however, I find that central-bank coordination and, to a lesser extent, large interventions increases the probability of success.