Key Macroeconomic Variables under Exchange Rate Volatility: Time-Varying Causality in the Presence of Structural Breaks and Nonlinearity

This paper explores the causal relationship running from exchange rate volatility to three macroeconomic variables in the case of Turkey. To that end, we first apply the classical Granger causality test introduced by Toda and Yamamoto (1995). We also use the time-varying Granger causality test developed by Shi, Hurn, and Phillips (2020) within the lag-augmented VAR model in the presence of empirically documented structural breaks and nonlinearities. A clear pattern that can be drawn from the causality results is that the causal channel from volatility to inflation is more sustained than causality from volatility to real GDP irrespective of size of the windows and selected recursive estimation algorithms. Besides, the causal channel from volatility to inflation coincides with time periods in which Turkey exhibits political and economic policy changes and suffers from increasing economic uncertainties during financial crises. The CBRT must strictly adhere to the CBRT Law and maintain its independence in order to ensure price stability as the unconventional monetary policy dictated to the bank by the government is itself the source of inflation. Finally, exchange rate volatility does not have predictive power for interest rates over the entire sample since the CBRT uses its foreign exchange reserves to offset the adverse effects of unexpected exchange rate shocks.