Show simple item record

dc.contributor.advisorBetancourt, Rogeren_US
dc.contributor.advisorPrucha, Ingmar Ren_US
dc.contributor.advisorKelejian, Harry Hen_US
dc.contributor.advisorEvans, Williamen_US
dc.contributor.authorKapoor, Muditen_US
dc.date.accessioned2004-05-31T19:56:10Z
dc.date.available2004-05-31T19:56:10Z
dc.date.issued2003-12-08en_US
dc.identifier.urihttp://hdl.handle.net/1903/143
dc.description.abstractThe first part of my dissertation considers the estimation of a panel data model with error components that are both spatially and time-wise correlated. The dissertation combines widely used model for spatial correlation (Cliff and Ord (1973, 1981)) with the classical error component panel data model. I introduce generalizations of the generalized moments (GM) procedure suggested in Kelejian and Prucha (1999) for estimating the spatial autoregressive parameter in case of a single cross section. I then use those estimators to define feasible generalized least squares (GLS) procedures for the regression parameters. I give formal large sample results concerning the consistency of the proposed GM procedures, as well as the consistency and asymptotic normality of the proposed feasible GLS procedures. The new estimators remain computationally feasible even in large samples. The second part of my dissertation employs a Cliff-Ord-type model to empirically estimate the nature and extent of price competition in the US wholesale gasoline industry. I use data on average weekly wholesale gasoline price for 289 terminals (distribution facilities) in the US. Data on demand factors, cost factors and market structure that affect price are also used. I consider two time periods, a high demand period (August 1999) and a low demand period (January 2000). I find a high level of competition in prices between neighboring terminals. In particular, price in one terminal is significantly and positively correlated to the price of its neighboring terminal. Moreover, I find this to be much higher during the low demand period, as compared to the high demand period. In contrast to previous work, I include for each terminal the characteristics of the marginal customer by controlling for demand factors in the neighboring location. I find these demand factors to be important during period of high demand and insignificant during the low demand period. Furthermore, I have also considered spatial correlation in unobserved factors that affect price. I find it to be high and significant only during the low demand period. Not correcting for it leads to incorrect inferences regarding exogenous explanatory variables.en_US
dc.format.extent962409 bytes
dc.format.extent1077054 bytes
dc.format.mimetypeapplication/pdf
dc.format.mimetypeapplication/pdf
dc.language.isoen_US
dc.titlePanel Data Models with Spatial Correlation: Estimation Theory and Empirical Investigation of the US Wholesale Gasoline Industryen_US
dc.typeDissertationen_US
dc.relation.isAvailableAtDigital Repository at the University of Marylanden_US
dc.relation.isAvailableAtUniversity of Maryland (College Park, Md.)en_US
dc.contributor.departmentEconomicsen_US
dc.subject.pqcontrolledEconomics, Generalen_US


Files in this item

Thumbnail
Thumbnail

This item appears in the following Collection(s)

Show simple item record