دراسة مقارنة لبعض طرائق تقدير الانموذج الديناميكي المكاني الخاصة بـ (Panel Data) مع تطبيق عملي == Compared Study To Some of The Methods of Estimating The Dynamic Spatial Model For (Panel Data) With Practical Application
Author name:
سهاد علي شهيد مجيد التميمي
Supervisor name:
حامد سعد نور الشمرتي
General topic:
Administration and Economics
Specific topic:
Statistics
Degree:
Doctorate
University:
Mustansiriyah University
Language:
Arabic
University location:
Baghdad
First pages:
07T4253 - p.pdf
Abstract:
يعد انموذج القياس الاقتصادي المكاني (Spatial Econometrics Model) احد انواع النماذج القياسية اذ تهتم بدراسة تاثير التفاعلات المكانية مابين الوحدات الجغرافية على الظاهرة المدروسة، تلك الوحدات من الممكن ان تكون مدن، بلديات، مناطق، دول، محافظات،... الخ. وت | Spatial Econometrics models are part of the Econometrics models as concerned with the impact of spatial interactions between geographic units. These units could be cities, municipalities, regions, states, provinces, etc..... It can also be used such models to explain the behavior of economic factors, geographic units, such as individuals, companies, or governments, if they were linked with each other through a network, but this type of studies in spite of a lot of use but less common.Should be noted that the spatial modeling studies differs from the time - series modeling studies, as the time - series studies focuses on the dependency between observation when a period of time and used the symbol (t - 1) to refer to the Lagged variables in time. It must be emphasized here that the (Spatial Econometrics) is not a direct extension for (Time - Series Econometrics) for two reasons; The first is that any geographical units can effect each other mutually, while it cannot be if they were tow observation at a certain time. Another factor most complex wide variety of units of measurement, which qualifies them for modeling spatial dependency (convergence, distances, bonding, etc...) compared with the measurement time dependency (time).In this thesis was to review some of the parametric techniques to estimate (Spatial Dynamic Panel Data Model (SDPD)), as there is dependent variable Lagged in time and extra (Endogenous Variables) in Spatial Dynamic Panel Data Model (SDPD) invalidate the use of estimation methods known like (Ordinary Least Square (OLS)) and the method (Maximum Likelihood (ML(. To overcome these problems, there are many empirical studies that have been applied in a manner estimate (Generalized Moment Method (GMM)), which were developed by each of the (Arellano & Bover 1995) and (Blundell & Bond 1998). As the main reason behind the use of estimator (GMM) to estimate the (SDPD) model is its ability to correct the interference (Endogeneity), which came by spatial lag variable also reason is due to the internal variables. In addition, it is estimated (GMM) is robust to some econometric issues such as measurement errors and poor (Instrumental Variable (IV)), it can also control the problem (Heteroskedasticity) and the problem (Autocorrelation) in the random error term, add to that the flexibility in the implementation of the calculations. In addition to the method (GMM) can be used (Quasi - Maximum Likelihood Estimation (QMLE)) when (Random Error) is distributed (Independent and Identical Distribution (iid)) with Mean zero and variance (?_?^2(.Estimated which is obtained be biased and inconsistent so it was (Bias Correction) to estimator (QMLE), also called the method of estimation of these (The Direct Approach).It has also been used (The Transformation Approach) for the purpose of processing the Spatial before the estimate to reduce bias as much as possible.As one of the most important goals is to study the effect of( Spatial Dynamic Effect)) sense of how dependent variable coefficients change that represents the (financial allocations for the program of economic sectors of the regions development) Lagged in spatially and in time and represented by )W_n Y_(n,t - 1)(between the provinces as well as the impact study of by Lag spatial dependent variable )W_n Y_nt( for each province of the Iraqi provinces and at certain period of time. Add to describe the effects of explanatory variables (Exogenous Variables) any (of economic sectors of each province for a period of a certain time) represented (b agricultural sector, industrial sector, transport sector, buildings and services sector, education sector) level of each province for the period (2008 - 2015).In the case of the first model any presence influences the spatial fixed only the better way transfers the larger time periods (T), the estimators (? ?_nT) be consistent (Consistence) and possesses a good estimator convergent properties correct the bias is more effective and has a proximity naturally parameter characteristics (?). The second model in the case of any effects the presence of spatial and temporal fixed together, was also the best way transfers from the direct method the larger time periods (T).As in the case of the practical side show in the case of both models weakness autoregressive spatial coefficient of investment allocations it does not have any interaction between the provinces where positively affecting the distribution of assignments better and to meet the economic development of each province
