Abstract
This dissertation consists of three chapters, each delving into various extensions of the New Keynesian model. The first chapter introduces a modification of the New Keynesian model to incorporate habit persistence and finite planning horizons, inspired by the work of Woodford (2019). This model, termed the Habit Persistence and Finite Horizon New Keynesian (HFNK) model, is shown to be empirically more robust and useful for monetary policy considerations than its standard, myopic, or finite-horizon counterparts. Our model reveals that planning horizons for households and firms are longer than typically suggested in the literature, under two years. The forecasting power of our model is found to outperform professional forecasts, particularly for output, showcasing its potential for real-world applications.
The second chapter expands upon the open economy New Keynesian model, incorporating agents with bounded rationality following \cite{G2020}. This development successfully addresses several puzzles related to the relationship between exchange rates and interest rates. It is found that employing actual exchange rate expectations, rather than rational ones, can make the uncovered interest parity (UIP) puzzles disappear. The model demonstrates that bounded rationality lessens the impacts of current monetary shocks and reduces the efficacy of forward guidance, particularly in more open economies. It also suggests that bounded rationality can enhance monetary spillovers, extend the persistence of the real exchange rate and net foreign assets, and exacerbate the unit root problem in small open economies.
The final chapter continues the exploration of bounded rationality within the open economy New Keynesian framework, offering a more behaviourally focused approach. This model again proves successful in addressing puzzling exchange rate and interest rate dynamics, and supports the idea of using actual, rather than rational, exchange rate expectations. This extension also brings into question the desirability of using the exchange rate as a nominal anchor in the context of optimal monetary policy. Interestingly, a welfare analysis conducted within this framework suggests that when agents fully and correctly anticipate the future, it results in more significant welfare losses than when they behave in a more adaptive and less fully informed manner. This counter-intuitive finding opens the door for further investigation into the role of behavioural factors in economic modelling.