What do robust policies look like for open economy inflation targeters?

For policymakers, thinking about best practice monetary policy means thinking about uncertainty. Open economy monetary policymakers face an additional source of uncertainty – exchange rate dynamics. This paper identifies policy rules robust to the open economy inflation targeters face in practice. For Knight (1921), uncertainty differs from risk because the policymaker does not know the nature of the uncertainty and is unable to form a probability distribution or risk statement, over different possible models. Hansen and Sargent (2004) apply Knight’s (1921) philosophy to the linear-quadratic control framework, recognizing that policy-makers work with models which are approximations to some true, unknown model and seek a rule that is robust to models close to the policymaker’s best approximation. While there exist some open economy robust control policy experiments (Leitemo and Söderström (2004) obtain analytical robust control solutions for a purely forward-looking new Keynesian model) the majority of the literature focuses on the closed economy. This paper calibrates a single open economy model to capture the key open economy dynamics for Australia, Canada and New Zealand, three of the earliest inflation targeters that form a useful dataset for identifying robust monetary policy rules in practice. Robust policies are found to respond more aggressively to not only inflation and the output gap, but also the exchange rate and its associated shock. This result generalizes to the context of a flexible inflation targeting central bank that cares about the volatility of the real exchange rate. However, when the central bank places only a small weight on interest rate smoothing and fears misspecification in only exchange rate determination, a more aggressive response to the lag of the exchange rate is not warranted. It is shown that the benefits of an exchange rate channel far outweigh the concomitant costs of uncertain exchange rate determination.