Economia comportamental e puzzles do mercado de ações: uma abordagem baseada em agentes

Abstract

Economic Sciences came across the 20th century with a paradigm shift in the explanation of market movements, motivated mainly by the explanatory inconsistencies for the real behavior of the investors when compared to the expected normative behavior. The expectation of rational and optimized behavior of economic agents was constantly questioned, especially in the 1980s and 1990s, when observed that investors, buyers and sellers acted not only by valuing optimized losses and gains but also making their decisions based in context, scenery, external influences, heuristics and influenced by emotions. This research uses an Agent-Based Approach to describe and explain some of the main anomalies or puzzles of the stock market, using a computational model that verifies the influence of fundamentalist and technical agents in an artificial market composed of fundamentalist agents with unlimited rationality and rational expectations, and by technical agents that have limited rationality and who adopt anchoring heuristic in the process of judgment and decision on the quantity of shares to be bought or sold. The results point to a significant influence of behavioral bias and heuristics incorporated into the model, supporting the occurrence of stylized facts – fat tails and volatility clusters – as consequences of behavior and decision making that is not based only on rationality or maximization of gains.