Estimating Central Bank Behavior in Emerging Markets: The Case of Turkey

Design of policy rules for an an emerging market central bank (EMCB) operating in an inflation-targeting framework presents additional challenges beyond those for describing the behavior of a central bank in a developed economy. Even though an inflation-targeting regime entails abolishing the exchange rate target in favor of an inflation target, it is more difficult for an EMBC to ignore movements in exchange rates given the relatively shallow depth of financial markets and the the high degree of dollarization. Additionally the EMCB may be forced to change the pursued exchange rate regime following a capital account reversal so that linear Taylor rules may be inadequate for describing EMCB reactions. We develop an empirical framework that addresses these issues and propose a new methodology to estimate unobserved variables such as expected inflation and potential output within the rule. Specifically, we employ a structural, nonlinear Kalman filter algorithm to estimate time-dependent parameters and unobserved variables, and we experiment with various exchange rate mechanisms that can be employed by an EMCB. This approach allows us to track any changes in EMCB behavior - including regime shifts - following a switch to inflation targeting. Using post-2001 data from Turkey, which is a fairly dollarized small open economy, we document that the Central Bank of Turkey has given relatively more importance to the inflation gap than to the output gap or to exchange rates, but not until some time after it had switched to an inflation-targeting framework.