Generalized Multiplicative Error Models

Abstract

In this chapter, we present generalizations of the basic multiplicative error model as introduced in Chap. 5. Section 6.1 discusses a class of ACD models which can be presented in terms of a generalized polynomial random coefficient model according to Carrasco and Chen (2002). We illustrate various special cases, discuss the theoretical properties and show empirical illustrations. In Sect. 6.2, we consider regime-switching ACD models allowing for parameters which might change in dependence of observable or unobservable characteristics. We concentrate on threshold ACD models, smooth transition ACD models as well as Markov Switching ACD specifications. Section 6.3 focuses on ACD models accommodating long range dependence in the data. In this context, we discuss different possibilities to capture long memory.