Monetary Policy Under Uncertainty

In this chapter, we take into account that real decision-making has to cope with control errors, imperfect information, and robustness issues. Intuition suggests that central banks should act less aggressively as a response to uncertainty, with policy response depending on the precision of noisy signals. The more precise the signal, the more active should be the response with Bayesian updating in case of a quadratic loss function. Similarly, actions should be dampened if there is multiplicative uncertainty about the transmission mechanism. Robust control theory, however, using minimax rules, suggests that under some conditions central banks might act more aggressively, aiming to avoid really bad extreme case outcomes. Optimal monetary policy strongly depends on variables which are hard to observe in reality. We show the challenges involved in estimating potential output, the natural rate of interest, and measuring expected inflation. We illustrate that simple rules, like the Taylor principle, may lead to quite ambiguous predictions if we take that issue seriously into account. Finally, we look at the impact of transparency and independence for central banks. Transparent communication is an essential ingredient of modern forward-looking policy. However, we show that there may be conditions when too much transparency might be harmful: Since public information plays also a coordinating role, private agents might overemphasize public signals.