ABSTRACT
This chapter shows the relationship between the dependent variable and a set of explanatory variables when the regression coefficients are purely random, and when they vary sequentially. It considers the estimation problem of a linear regression model first when the intercept in the model is varying sequentially and then when both the intercept and slope coefficients are varying. An advantage of the Bayesian estimation is that one can incorporate the prior information about. The coefficient levels in the adaptive process are nonstationary; therefore likelihood functions for them cannot be specified. The chapter discusses on a study of the Phillips curve with the purely random coefficients. It concludes that modelling wage inflation with random coefficients generates more precise forecasts than modelling with fixed coefficients. Furthermore, the wage inflation equation with stochastic coefficients forecast rather well, which suggests that the shift in the Phillips curve for the years following 1966 probably did not take place.
