PRINCIPAL COMPONENT ANALYSIS TO DEVELOP PERFORMANCE INDEXES IN THE PUBLIC SECTOR

Abstract

This investigation has as its main goal to clarify the necessary procedures to develop performance indexes in the public sector with the use of the Principal Component Analysis. As specific goals: 1) to identify the requirements of the evaluative research in the public sector; 2) To list the necessary assumptions to use the Principal Component Analysis; 3) to develop an empirical model of the calculation of the performance index conducted by the Principal Component Analysis; and 4) to show from a practical example how to develop performance indexes in the public sector. As a practical example, it was developed the Social performance Index of the States and the Federal District of 2013, based on the per capita social functions: social assistance; social security; health; education; culture; citizenship’s rights; urban planning; housing; sanitation and sports and entertainment. The financial data about the government functions were extracted from the website of the National Treasury Bureau (STN). The Federal District exhibited the best performance and the State of Alagoas, the worst performance. The guiding hypothesis of the investigation that there will be greater spending in the social functions in the states with greater income inequality could not be proved statistically, but the study highlights lack of administrative efficiency in the allocation of social resources.