بررسي رابطه بين رشد اقتصادي و توزيع درآمد در ايران طي دوره 1393-1350 رهيافت رگرسيون غيرخطي انتقال ملايم

(The Impact of Economic Growth on Iranian Income Distribution, (Nonlinear LSTAR approach

هدف اصلي اين مطالعه بررسي ارتبـاط بين متغيرهاي رشـد اقتصادي و تـوزيـع درآمد، طي سال هاي 1350-1393 در اقتصاد ايران مي باشد. بر اساس مباني نظري، بين رشد اقتصادي و توزيع درآمد رابطه U معكوس وجود دارد كه اين ارتباط، به فرضيه سيمون كوزنتس در ادبيات اقتصاد كلان معروف شده است. براي اين منظور، از شاخص ضريب جيني به عنوان شاخص اندازه گيري توزيع درآمد، تغييرات توليد ناخالص داخلي ايران به عنوان شاخص اندازه گيري رشد اقتصادي و رويكرد LSTAR يا روش انتقال ملايم خودرگرسيوني لوجستيكي، به عنوان رويكرد اقتصادسنجي استفاده و مدل اقتصادسنجي با استفاده از نرم افزار JMulti تخمين زده شده است. نتايج اين مطالعه، ارتباط مستقيم و غيرخطي بين رشد اقتصادي و توزيع درآمد را در دوره مورد بررسي، در بخش خطي و غيرخطي مدل تاييد مي نمايد. همچنين بر اساس نتايج اين مطالعه، برآيند تاثير رشد اقتصادي دوره جاري و قبل بر توزيع درآمد دوره جاري مثبت و معني دار است. از اين رو مي توان گفت كه دولت در كنار اجراي مرحله دوم هدفمندي يارانه ها، با هدف كاهش فاصله طبقاتي، به عنوان يكي از اهداف اجراي آن، از اعمال سياست هاي مناسب به منظور دستيابي به رشد اقتصادي بالاتر و در نتيجه كاهش ضريب جيني نبايد غافل شود.

Some subjects such as economic growth، optimum distribution of national income and also economic justice have been considered by the economics and economic policies in Iran. In this regard، there are a variety of ideas about the interaction effects of distribution of incomes and eeconomic growth. According of the foundations of theoretical، there is a U reverse relation between distribution of national income and economic growth. This curve in known as Simon Kuznets hypothesis in economic literature of growth. Some different result have been obtained about the relation between economic growth and income inequality based on several studies in Iran and out of Iran. According to doing the law enforcement targeted subsidies in Iran during recent years، one of the most important goal of this plan was to improve distribution of national income among the poor layers of social. In this regard، this study aims to evaluate the relationship between economic growth and income distribution during 1350-1389 in Iran. For this purpose، Gini index for income distribution measurement and autoregressive logistical approach or methodology smooth transition logistic smooth transition auto regression (LSTAR) is used as the econometric approach. In this method، the relation between variables is changed nonlinearly. So، first the problem is modeled nonlinear. According to the number of observations that are less than 100 observations، Schwarz- Bayesian criterion is used to in order to finding the optimum amount of lag. Based on this criterion، the optimum amount for both dependent and independent variable is 1. Based on result about estimation the model and rejection two hypothesizes and and، the model of LSTAR1 is selected. Next results show that gross domestic product (GDP) growth has a negative and significant effect on Gini index as a criterion for measurement income distribution both in linear and nonlinear cases in the period under review. Non-linear relationship between these two variables shows that the optimal level of the Economic Growth involves maximizing Gini coefficient in Iran. On the other hand، more increasing in Gini index means shows more unequally in income distribution. Therefore adoption and deploying polities for increasing growth economic will cause decreasing Gini index and more equally income distribution in Irani economic. This result also show that، the hypothesis of Kuznets about existing a nonlinear relation between economic growth and income distribution cannot be rejected. In order to evaluate the nonlinear estimated model some tests were used such as ARCH- LM test، lack of autocorrelation، and stability of parameters. This evaluation was done by using the JMULTI software. The results of evaluation confirm that there is no error relevant to existing autocorrelation in none of lags. This result show that there is no autocorrelation in variable lag of the gross domestic product (GDP) growth in Iran economic. Additional analysis show that there is no other nonlinear variable in the estimated model and so، the null hypothesis about suitable specification of the model cannot be rejected. Also in order to evaluate the dissimilarity of variance between disturbances of the estimated phrases، the Auto Regressive Conditional Heteroskedasticity Lagrange Multiplier (ARCH-LM) test was done. The result of analysis in this section presents that the hypothesis of absence of dissimilarity of variance between disturbances of the estimated phrases، cannot be rejected. Additional analysis confirm that equality of the coefficient in linear and nonlinear model is rejected. Finally the Jarque-Bera test confirm that the disturbance phrases have a normal distribution. The basic results of this study have a good compatibility with foundations of theoretical and also the result of studies of Uge Panizza (1999)، Rafael Gomez & David K. Foot (2003). Angles – castro، Gerardo (2006)، and Ben – David Nissim (2007