Views expressed are our own and do not necessarily reect those of the DNB. We are grateful to

We study the implications of uncertainty for ination targeting. We apply Brainards static framework which assumes multiplicative uncertainty in the monetary transmission. Brainards main result is that in the presence of un-certainty, monetary authorities become naturally more cautious. But this also implies that monetary objectives are seldom achieved. We therefore attempt to nd a monetary rule that reaches the objectives set more often and improves the welfare of the Central Bank. Such a rule is the result of a new algorithm that we put forward, in which the ination target is state contingent. The Central Bank sets therefore (as an auxiliary step), a variable ination target that depends on both the degree of uncertainty as well as the shocks that occur each time. If the bene\u85ts of reaching the ination target are properly accounted for in the loss function, we show that such an optimisation proce-dure helps the CB attain its objectives more often, thereby reducing the losses incurred. Moreover, and as a corollary to such an approach, the rule derived is ex ante neutral to the degree of uncertainty