We reconsider Svensson's inflation-targeting proposal in a model where the need to raise seigniorage revenues determines the socially optimal inflation rate and distortionary taxes cause the inflation bias. Interpreting the targets as contracts, we show that the interaction between fiscal and monetary policy complicates the structure of the optimal contract. Moreover, if the commitment technology is imperfect, "highish" targets generate lower inflation than targets, which are too low to be credible. Alternatively, interpreting inflation targets as policy delegation to a non-distortionary target-conservative agent, we show that target-conservative bankers are public-expenditures conservative. Unfortunately, only idiosyncratic views about the benefits from public expenditures can be invoked to justify expenditures-conservatism, implying that target-conservative agents are also weight-conservative.