Analysis of Turkey's Domestic Debt Stock Increment with a Robust Partial Least Squares Regression Method: RSIMPLS and a Robust Principal Component Regression Method

Abstract

The aim of this study, different from other studies about Turkey's domestic debt, two popular biased robust methods Robust Principal Component Regression (RPCR) and RSIMPLS (a Robust Partial Least Squares Regression method) are used for modelling domestic debt increment in Turkey in existence of both multicollinearity and outlier in the data set. These methods have been compared in terms of predictive ability and goodness of fit by using a robust Root Mean Squared Error of Cross-validation (R-RMSECV) and a robust Root Mean Squared Error (R-RMSE) statistics, respectively. The analyses results showed that in order to decrease the internal debt stock in Turkey, firstly, internal debt service (IDS) must be decreased. Because in developing countries such as Turkey chronicle public deficit is a most argued economical problem, IDS is not directly shows the country’s domestic debt payment power and debt payments are done by new borrowings so IDS creates a new borrowing requirement, therefore, increase the internal debt stock. Secondly, especially public sector borrowing requirement-PSBR (means all the deficits in public sector such as social security institutions, local authorities etc.) and exchange rate must be decreased.