This thesis provides an extension of panel data models on the analysis of Economics and Finance Data, discusses methods of estimation and evaluation for such models and presents empirical applications. The thesis consists in three Chapters. The first Chapter proposes three alternative approaches to test the Permanent Income Hypothesis (PIH) in the context of dynamic panels: the aggregate consumption approach, the Euler equation approach and finally Friedman (1957)’s original characteristic tests. The empirical evidence, using the British Household Panel Survey (BHPS) data, strongly supports the PIH. Hence, the analysis presented can be considered as supporting the view that empirical tests of PIH, based on aggregate time-series data, might suffer from misspecification or overlook some fundamental characteristics of micro data. The second Chapter addresses the issue of testing for factor price misspecification via a panel data approach. A theoretically coherent framework based on panel data techniques has been constructed. This allows for both the homogeneous and heterogeneous parameters that are present when testing for anomalies in factor pricing models. The tests presented have a clear advantage over the traditional two-pass based tests because they do not suffer from the errors in variable problem and have all the usual desirable asymptotic properties associated with the maximum likelihood approach. The empirical illustration shows that book to market equity and market value firm specific characteristics help explain asset returns in the UK over 1968-2002 even when all three of Fama and French’s factors are present. This finding, which is in contrast to much of the literature, may be due to the improved efficiency of our estimates and power of our tests relative to those based on the two-pass method predominant in the existing literature. Finally, we find that the overall significance of firm size is attributable to its importance during the 1980’s subsample, a period of relatively calm stock market growth. Lastly the third Chapter presents an application of Hausman-Taylor estimation in heterogeneous panels with time-specific common factors to gravity models of intra-EU trade. As an extension to the Hausman-Taylor procedure, an efficient instrumental variable estimation of a panel data, which includes time-specific common factors and their heterogeneous individual parameters, is presented. The underlying econometric techniques are developed and an alternative source of instruments is suggested. This methodology is applied to gravity models for international flows of trade using data on fifteen European countries over 42 years (1960-2001). Following the most recent developments of the literature, a complete analysis of the sources of bilateral trade amongst European countries is presented using three different specifications. The final model is also originally extended in order to allow for observed common factors and their heterogeneous parameters. The empirical evidence confirms the effectiveness of the gravity model in explaining international trade flows. Results also encourage the use of our extended approach as a valid alternative to the basic time dummy specification.
Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:661747 |
Date | January 2004 |
Creators | Serlenga, Laura |
Publisher | University of Edinburgh |
Source Sets | Ethos UK |
Detected Language | English |
Type | Electronic Thesis or Dissertation |
Source | http://hdl.handle.net/1842/25172 |
Page generated in 0.0023 seconds