Quantitative modeling in international finance: implications of heterogeneous investors, artificial neural learning and frictions

Three topics in international finance are considered.
The first topic deals with international investment strategies, which consider investors with different objective functions holding tradeable and non-tradeable assets and liabilities, stochastic investment opportunity sets and Poisson events in price processes. The balance sheet is crucial in determining optimal investment strategies.
The second topic deals with artificial neural networks in forecasting. Based upon fundamental variables of a theoretical exchange rate model, econometric models and neural networks predict the UK£/US$ and the Yen/US$ exchange rates. In this fundamental setting, neural networks on average perform best at longer horizons.
The third topic deals with optimal switching between two riskless assets in presence of stochastic investment opportunity sets and transaction costs. When the instantaneously riskless interest rate difference follows a random walk without drift and traders receive a performance dependent compensation, small transaction costs introduce important inactivity in trading.

Publication infos:
Genève, Institut universitaire de hautes études internationales, 1996
Publication year:
Number of pages:
156 p.
PhD Director(s):
Directeur de thèse: Professeur Hans Genberg
Call number:

 Record created 2011-06-03, last modified 2019-09-30

Rate this document:

Rate this document:
(Not yet reviewed)