Please use this identifier to cite or link to this item:
Dynamic multivariate analysis of a small open economy: The case of Hawai'i
|uhm phd 4375 r.pdf||Version for non-UH users. Copying/Printing is not permitted||5.47 MB||Adobe PDF||View/Open|
|uhm phd 4375 uh.pdf||Version for UH users||5.47 MB||Adobe PDF||View/Open|
|Title:||Dynamic multivariate analysis of a small open economy: The case of Hawai'i|
|Contributors:||Bonham, Carl S (advisor)|
show 5 moreEconomics
|Date Issued:||Aug 2003|
|Publisher:||University of Hawaii at Manoa|
|Citation:||Zhou, Ting (2003) Dynamic multivariate analysis of a small open economy: The case of Hawai'i. Ph.D. dissertation, University of Hawai'i, United States -- Hawaii.|
|Abstract:||The main objective of the dissertation is to apply recent advances in modern econometric analysis, namely cointegrating Vector Autoregression (VAR) and Bayesian VAR (BVAR) to a small open regional economy like Hawaii. This is accomplished in three related yet independent essays demonstrating how regional modeling and forecasting can benefit from these latest developments. The first essay concentrates on the cointegrating VAR analysis, applying it to Hawaii's premier industry-tourism. Recent research in the literature on identified cointegrating VARs emphasizes the need to rely on economic theory to impose weak exogeneity assumptions, guide the search for long-run just (over) identifying restrictions and shrink the model to the most parsimonious representation. While cointegration analysis has gradually appeared in the empirical tourism literature, the focus has been exclusively on the demand side with no use of the latest identification techniques. A complete Hawaii tourism model is developed, exploiting Hall, Henry, and Greenslade's (2002) theory-directed sequential reduction methodology. Both demand and supply factors are emphasized in identifying long-run cointegrating relationships. The second essay applies the BVAR methodology to another key sector in regional modeling-construction. This essay represents the first application of priors on linear combinations of parameters-namely, sums of coefficients and dummy initial observation priors - in a BVAR construction forecasting model. I find that including these priors does not necessarily improve forecast accuracy at medium to long horizons, especially when the series are integrated and there is more than one cointegrating relationship. The third essay extends the second essay to deal with the entire regional economy. All regional models must deal with the inavailability of expenditure data at the state and local levels. This problem typically leads researchers to use either a single highly restricted VAR, or BVAR, or a model of pseudo theory driven equations. In contrast, my third essay makes use of BVAR blocks to model proxies for the expenditure categories in a traditional macro structure. Compared with existing regional BVAR models, the current setup is more complete in accounting for both the intra-action of sectors within the region and the inter-action of the region with external drivers.|
|Rights:||All UHM dissertations and theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission from the copyright owner.|
|Appears in Collections:||
Ph.D. - Economics|
Please email email@example.com if you need this content in ADA-compliant format.
Items in ScholarSpace are protected by copyright, with all rights reserved, unless otherwise indicated.