The approach to dynamic modeling with latent variables has been developed on the base of H. Wold's partial least squares (PLS). The original PLS estimation algorithm is virtually applicable. In addition to that lagged and leaded latent variables are used in the iterative process of estimating the weights of the manifest variables. The path coefficients are estimated by OLS. A redundancy coefficient allows to measure the forecasting validity. Finally the algorithm has been programmed in XploRe .
This chapter surveys the theoretical background and explains how
dynamic partial least squares models are implemented in
XploRe
.
The last part focuses on an example for German share prices.