Title :
Modelling risk premium of repo interest rate in the SSE
Author_Institution :
Sch. of Manage., Fudan Univ., Shanghai, China
Abstract :
With weekly data of repo rates in the SSE, it is found that the expectations hypothesis fails to explain the repo rates, and risk premiums are significant and time-varying. One-factor and two-factor Guassian essential affine models are estimated to model the time-varying risk premiums, and the likelihood ratio test shows that the two-factor model has no significant improvement over the one-factor model. Although one-factor Guassian essential affine model fits the risk premiums very well, the model doesn´t fit the means and standard deviations of the repo rates.
Keywords :
Gaussian processes; economic indicators; maximum likelihood estimation; modelling; risk analysis; stock markets; SSE; Shanghai Stock Exchange; likelihood ratio test; one-factor Guassian essential affine model; repo interest rate; time-varying risk premium modelling; two-factor model Guassian essential affine model; Cybernetics; Economic indicators; Instruments; Machine learning; Macroeconomics; Pricing; Risk management; Security; Stock markets; Testing; Interest rate; Kalman filter; affine model; risk premium;
Conference_Titel :
Machine Learning and Cybernetics, 2005. Proceedings of 2005 International Conference on
Conference_Location :
Guangzhou, China
Print_ISBN :
0-7803-9091-1
DOI :
10.1109/ICMLC.2005.1527541