Record:   Prev Next
作者 Nocetti, Diego Carlos
書名 Essays in macroeconomics and behavioral economics
國際標準書號 9780542648717
book jacket
說明 73 p
附註 Source: Dissertation Abstracts International, Volume: 67-04, Section: A, page: 1470
Adviser: William T. Smith
Thesis (Ph.D.)--The University of Memphis, 2006
My dissertation consists of three essays that analyze diverse economic implications of uncertainty in policy making and in individual decision making. In the first essay I describe how decision makers form forecasts in the face of model uncertainty and the implications of model uncertainty for the performance of pooled forecasts ex-post. I show that the introduction of model uncertainty into the formation of expectations of optimizing agents can account for the stylized fact that pooled forecasts frequently outperform individual forecasts of economic time series. In the second essay I study the interaction between limited cognitive resources and information acquisition. I consider how investors allocate mental effort to learn about the dividend process of a number of assets, retrieving information from a stock of memories. As a result, parameter uncertainty arises endogenously. I show that this endogenous estimation risk can help explain a number of stylized facts in financial markets, including the risk premium and excess volatility puzzles, the higher volatility of less familiar equities, and the predictability of stock returns. In the third essay I develop and test the hypothesis that uncertainty about the solvency of a Central Bank might lead to currency and banking crises. I develop an operational model to calculate Central Bank's value at risk and illustrate the methodology using data from the recent financial crisis in Argentina. I show that the approach can provide higher accuracy and announce the probability of a crisis sooner than other indicators
School code: 1194
DDC
Host Item Dissertation Abstracts International 67-04A
主題 Economics, Finance
Economics, Theory
0508
0511
Alt Author The University of Memphis
Record:   Prev Next