Attenuated or Augmented? Monetary Policy Actions under Uncertain Economic State and Learning

Monetary policy is conducted in an environment of considerable uncertainty. In particular, Bernanke (2007) emphasizes that monetary authority faces substantial challenges in determining the sources of variation in macroeconomic variables. What can policy-makers do when supply and demand disturbances are unobserved and indistinguishable? Is a policy rule widely conducted otherwise still appropriate? I explore these questions by developing a novel New Keynesian DSGE model featuring uncertain economic state characterized by the unobservability and indifferentiability between supply-side total factor productivity (TFP) shock and demand-side investment specific technology (IST) shock. Private agents and policy-makers have access to noisy signals about these two productivities, learn from those signals, and make decisions based on subjective expectations. Given the estimated model, I show that, compared to the environment without uncertainty, economic agents behave similarly if shocks can be observed or quickly recognized, but differently if shocks take time to be revealed. Furthermore, a policy following the Taylor rule, which responds to inflation and output gap, still has the power to accommodate the TFP shock and offset the IST shock, even though these shocks are not observed or differentiable. Finally, the uncertainty generates different welfare implications for monetary policies: i) the optimal policy, despite minimizing the aggregate volatility and thus welfare losses, can achieve only a constrained effi cient state associated with an ineffi ciently low output level, and ii) Compared to the Taylor rule, the inflation targeting rule generates higher welfare under uncertainty, but lower welfare under certainty. This implies that policy actions should be adjusted in accordance with different economic environments.