City Research Online

Abstract

The evaluation of the impact of the news effects is one of the key questions in financial economics and a hot topic in recent studies of macroeconomic analysis. It may not be the act of releasing information to the market which is important, nor the (gross) information embodied in the estimate itself, rather, it is the extent to which the actual announcement differs from the expected which determines the response of the market to the new information (Kim et al. 2004). The aim of this thesis is to increase the knowledge of the impact of macro news, coming from scheduled macro announcements, on the US interest rates term structure. Chapter 1 reviews the definition and scope of the news concept and introduce a description of the data and its availability. Also, it presents the literature on the impact of news over stocks, foreign exchange markets and mainly interest rate term structure. The typical econometric weakness to address is that the literature usually starts the model selection from a limited General Unrestricted Model (GUM) that only considers a pre-selected pool of economic indicators. Given the myriad of economic announcements the researcher ends up dealing with a considerable model size that makes difficult the model selection procedure. Specially, the question on how to reduce the GUM to the Local Data Generation Process (DGP) raises naturally. In general, the amount of exogenous variables (announcements from the economic calendar) clearly points out to the impossibility to pick up easily a single model from the GUM. Moreover, the literature never studies the effects of all the possible economic announcements as the complexity of the models limited the analysis. This thesis adopts automatic selection devices such as PcGets to avoid pre-selection biases. The use of model selection techniques constitutes an important evolution in the way the literature handles the large number of economic indicators, without limiting the analysis to a subjective group of variables.

Publication Type:	Thesis (Doctoral)
Subjects:	H Social Sciences > HG Finance
Departments:	Bayes Business School Bayes Business School > Bayes Business School Doctoral Theses Doctoral Theses

Export

Downloads