Use the non-linear least squares method to estimate the regression model if there is a problem of heterogeneity of conditional variance (With practical application of cash presentation data)

Abstract

This research aims to construct a regression model with an dependent variable and independent variables for financial and economic time series data in case of existence of Autoregressive Conditional Heteroskedasticity) ARCH) , through the application of Non- Linear Feasible Generalized Least Squares (Non Linear – FGLS) . In order to achieve the aim of the research , a working mechanism has been performed , including (9) successive stages , taking into account the conditions and assumption required to construct the regression model , and the heterogeneous conditional variation model (ARCH) In a manner that could be relied upon to make estimates , This method has been applied to data from within the country represented by cash presentation data y_t , Assets and liabilities x_t1 ، x_t2 , Using the statistical packages (EViews 9 , STATISTICA , gretl ) .