Construction and forecasting performance of varying coefficient inflation models for the UK and China.
Although various theoretical and applied papers have appeared
in the field of Varying Coefficient (VC) modelling, little is
available in the literature on inflation. The aim of this thesis
is to evaluate the performance of the VC model and to fill the gap
in applications of the VC approach to inflation processes both in
the UK and China.
Statistical analyses suggest that the inflation processes are
unstable for both countries and the instabilities are mainly
associated with major economic and institutional changes. The
State Space (SS) model, in conjunction with the Kalman Filter
algorithm, is then introduced to simulate the structural change of
inflation models in the two different types of economies.
The performance of a number of existing constant coefficient
inflation models in the UK and that of their VC alternatives are
compared and the advantages of the VC approach are revealed.
A general VC inflation model for the UK, that nests both the
Keynesian and monetarist ideas and a VC inflation model based on
the excess demand hypothesis for China are developed. These VC
inflation models are considered to be more accurate and robust in
inflation forecasting because they take account of structural
instability in the inflation generating processes.
The empirical analysis gives results broadly consistent with
the VC models developed. This confirms the author's initial
assumption that the VC inflation specification will be superior to