Two Essays on Monetary Policy and Exchange Rates in Thailand

This dissertation presents two essays on monetary policy determination and exchange rate forecasting in Thailand. Thailand is a small open economy using an inflation-targeting framework with a managed floating exchange rate. The Bank of Thailand uses the policy interest rate as the main monetary policy instrument and intervenes in the exchange rate market only when necessary to prevent excess volatility and to achieve economic policy goals. Both interest rates and exchange rates shape the economic decisions of households and firms and thus play an important role in economic conditions and stability. Therefore, it is crucial to understand the determinants and dynamics of these variables.

The first paper investigates the Bank of Thailand's interest rate determination using the Taylor rule, which posits that the optimal interest rate is a function of inflation and the output gap. Because Thailand is a small open economy highly exposed to international trade and capital flows, I also include international factors (the federal funds rate and exchange rates) in the Taylor equations. I find that the Bank of Thailand significantly responds to expected inflation, which is consistent with an inflation-targeting framework, while the output gap does not play an important role in the monetary policy reaction function. I also find that the Bank of Thailand significantly responds to international factors, in particular adjusting the interest rate in the same direction as the federal funds rate and reacting to exchange rate movements.

The second paper evaluates out-of-sample exchange rate forecasting for the Thai baht vis-a-vis the U.S. dollar (USDTHB) using fundamental models from the literature such as the purchasing power parity model, the uncovered interest rate parity model, the monetary model, and the Taylor rule fundamental model at different forecast horizons and using various rolling window sizes. Using the Diebold-Mariano/West test, I find that the Taylor rule with inertia and asymmetric coefficients model and the monetary model with sticky prices have the strongest ability of the fundamental models to predict exchange rate movements. This Taylor rule model outperforms the random walk starting at the six-month forecast horizon, while this monetary model outperforms the random walk starting at the one-year forecast horizon. Then, I construct a hybrid model combining variables from these two top-performing models; pairwise comparison determines that this hybrid model outperforms both models individually. Additionally, I find that the exchange rate predictability of all models tends to improve as the forecast horizon increases, with the highest rejection against the random walk at the three-year forecast horizon while there is no significant evidence of short-run predictability.

Overall, I find that Thailand's monetary policy is best described by the Taylor rule, under which the Bank of Thailand strictly targets inflation in accordance with its economic goals. Additionally, the Bank of Thailand responds to changes in exchange rates and the federal funds rate. Therefore, based on this augmented Taylor rule, the nominal exchange rate can be forecasted using the Taylor rule with inertia and asymmetric coefficients model. However, exchange rate predictability is only found at the six-month horizon and beyond. I also find that by incorporating useful information from various macroeconomic variables, the hybrid model has the strongest predictive performance of all models.