MTV Model and its Applications

Abstract

As has been stated, many financial time series are of multi-variate covariational structure which evolves gradually and will not admit a profitable opportunity without risk. Hence this point must be taken into account in efficiently analyzing such volatile time series as stock prices, interest rates, exchange rates, etc. In particular, a multivariate approach is more appropriate to des-cribe these volatile phenomena and a proper length of sample period is required to be chosen. A typical multivariate time series model sometimes used in practice is VARMA (vector-valued autoregressive moving average) model. However, as we discussed in Chapter 4, the model has the two major difficulties in applications;

• (a) overflow of parameters

• (b) lack of identifiable models