آزمون ناهمساني واريانس و همبستگي فضايي در مدل رگرسيوني پانلي با مولفه هاي خطاي يك طرفه

چكيده انگليسي :

Testing for heteroskedasticity and spatial correlation in a random e ects panel data model Mostafa Gholizadeh M Gholizadeh@math iut ac ir 22 January 2013 Department of Mathematical Sciences Isfahan University of Technology Isfahan 84156 83111 Iran Supervisor Dr Soroush Alimoradi Salimora@cc iut ac ir Advisor Dr Iraj Kazemi ikazemi@sci ui ac ir 2010 MSC 62Kxx 62K05 62K10 Keywords Panel data Heteroskedasticity Spatial correlation Random e ects LM test AbstractThe combination of the time series data and the cross sectional data called panel data In panel data the information of multiple units N during a speci ed time period T are studied The regressionmodel which used to analyze these data known as a panel regression model in randomly drawn samples at the individual level one dose not usually worry about cross sectioncorrelation However when one starts looking at a cross section of countries regions states counties etc these aggregate units are likely to exhibit cross sectional correlation that has to be dealt with there is an extensitive literature using spatial statistics that deals with this type of correlation thisspatial dependence models are popular in regional science and urban economics more speci cally these models deal with spatial interaction spatial autocorrelation and spatial structure spatial het erogeneity primarily in cross section data spatial dependence models may use a metric of economicdistanc which provide cross sectional data with a structure similar to that provided by the time indexin time series so if the cross sectional units have the location these data called spatial panel data Each location may have di erent e ect on data because of the heterogenous nature of the spatiallocations These e ects may consider xed or random e ects As one of the most widely applicationof linear models with random e ects is used in panel data modelling so in this thesis we would liketo consider linear model with random e ects in the standard error component model assumes that the regression disturbances are homoskedasticwith the same variance across time and individuals this may be a restrictive assumption for panels where the cross sectional units may be of varying size and as a result may exhibit di erent variation Furthermore If the variance of all components be constants the panel regression model would callpanel regression model with standard error components But If some of the error components do nothave the same variance It is said that there is the heteroskedacity in the model in this thesis a paneldata regression model with heteroskedastic as well as spatially correlated disturbances is considered and a joint LM test for homoskedasticity and no spatial correlation is derived in addition a con ditional LM test for no spatial correlation given heteroskedasticity as well as a conditional LM testfor homoskedasticity given spatial correlation are also derived These LM tests are compared withmarginal LM test that ignore heterokedasticity in testing for spatial correlation or spatial correlationin testing for homoskedasticity

قلي زاده، مصطفي

آزمون ناهمساني واريانس و همبستگي فضايي در مدل رگرسيوني پانلي با مولفه هاي خطاي يك طرفه

داده هاي پانلي , اثرات تصادفي , آزمون LM