ENDOGENOUS FLUCTUATIONS IN MACROECONOMICS: THE ROLE OF HETEROGENEITY

My Ph.D. thesis contributes to the growing literature investigating the role of heterogeneity in the macroeconomics volatility. Particularly, it focuses on bounded rationality and heterogeneous expectations in financial accelerator frameworks. In the first paper, I present a literature review tackling the theoretical and the empirical challenges of bounded rationality and heterogeneous expectations. More precisely, I decided to analyse the Adaptive Belief System (ABS) because it is relevant to the purpose of solving the “wilderness problem” of bounded rationality. Moreover, it introduces complex dynamical evolution in the system and it can describe many stylized financial and macroeconomics phenomena, like fat tails, unpredictable returns and excess volatility. I conclude the survey showing some experimental analyses suggesting the importance of heterogeneity and bounded rationality in the expectations formation and in the evolution of the economic system. In the other two papers, I elaborate two financial accelerator (FA) frameworks focusing on the role of heterogeneous expectations, investment decisions and macroeconomic fluctuations. In the first one, starting from the Bernanke-Gertler-Gilchrist financial accelerator (1999), I develop an Agent-Based financial accelerator introducing bounded rationality, heterogeneous expectations, actual bankruptcy and a balance sheet for the financial intermediary. I specify a setup in which the heterogeneity is inserted in the agents’ wealth as well as in their heuristics. The agents make mistakes in forecasting future macroeconomic variables and update their beliefs when new information becomes available. Since they commit mistakes in their investment and borrowing decisions, the entrepreneurs may not be able to fulfil their debt and therefore go bankrupt. To account for the losses of defaults I introduce the balance sheet of the financial intermediary. Then, the bankruptcy affects the credit channel: firstly because the bank will settle an extra cost to the defaulted entrepreneur, secondly because banks with lower financial soundness will fix on average higher interest rate on loans. Finally, I explore the macroeconomic volatility running some simulations which consider different heuristics and different monetary policies. My results suggest that a monetary authority should take into account the “sentiment of the market” when designing its policy. Indeed, applying the same monetary policy may have different consequences on macroeconomic volatility when the expectations are non-identical. However, if two monetary policies are implemented in the same scenario, it seems that the strongest monetary policy will reduce the waves of optimism and pessimism better stabilizing the macroeconomic environment. In the second model, I propose a financial accelerator in which the evolution of expectations is based on the adaptive belief system. Within this framework, the entrepreneurs have cognitive limitations and are not able to forecast in advance the actual return on capital. However, when new information becomes available, they can compute their investment performance and switch to the most performing heuristic. Through this mechanism, they update their beliefs on future investment return introducing complex dynamics in the model. In the last part of the paper, I explore the macroeconomic volatility of the system considering different heuristics and monetary policies. One the one side, the core results suggest that no monetary policy is able to quickly stabilize the system completely; some fluctuations persist for many periods. Moreover, flexible inflation targeting policies yield lower fluctuations but these are more persistent. On the contrary, strict inflation targeting policies produce deeper macro-volatility but lower persistency. Finally, the stabilizing effect of the monetary policy strongly depends on the nature of the heuristics and, counterintuitively, the volatility is higher in scenarios with more sophisticated heuristics.