TITLE:
Averaged-Calibration-Length Prediction for Currency Exchange Rates by a Time-Dependent Vasicek Model
AUTHORS:
Tomasz Serafin, Anna Michalak, Łukasz Bielak, Agnieszka Wyłomańska
KEYWORDS:
Modeling, Vasicek Model, Time-Dependent Model, Long-Term Prediction, Calibration, Estimation
JOURNAL NAME:
Theoretical Economics Letters,
Vol.10 No.3,
June
16,
2020
ABSTRACT: The mining business is extremely sensitive to market
factors in price behavior. One of the main risk factors in the KGHM, one of the
biggest mining companies in the world, is the currency exchange rates prices.
Thus, one of the main problems from the market risk management perspective is
to properly predict the dynamics of the currency exchange rate data in the
long-term horizon. In this paper, we propose to model the data by the so-called
extended Vasicek model, which is a natural generalization of the classical
Vasicek model, also known as the Ornstein-Uhlenbeck process. The classical
model is very popular in the financial data modeling, however, it does not
capture the possible changes in the long-term mean and long-term variance. The
extended model takes into consideration the fact that the dynamics of the data may change over time by using time-varying
coefficients. Applying the extended Vasicek model, we demonstrate the problem
of long-term prediction and propose a new approach in this context which is
based on the averaging of the
predictions obtained from different calibration sample lengths.