Identification of Quality of VAT Data: The Case of Albania

Abstract

The increasing demand of users for quick and qualitative statistics, have increased the focus of statisticians from different countries to develop strategies, produce systems, compile statistical methods for the use of administrative sources as primary source. An important issue of official statistics is the reduction of the response burden from enterprises. Administrative sources are frequent in time, have wide scope coverage, are fast in terms of deadlines, but in content they carry suspicious values. In the case of Quarterly Survey these suspicious values affect the quarterly growth rate analysis. In Short Term Statistics cases where the value between the quarters is not accurate give a wrong picture of the development. In the Structural Survey, an annual net sale gives a wrong picture of the structure of the economy. Sources used for this paper are Value Added Tax and Quarterly Survey of Enterprises. The paper tries to identify the different patterns of the VAT in Albania by analysing two years VAT data, 2016-2017 and propose the editing method for correcting them. As the focus is the use of VAT in Short Term Statistics, the patterns are identified in the panel data created by linking with the same ID code the businesses on both sources. Various methods have been used from other countries’ statistics for the editing of these suspicions values. The experience of these countries is studied in orded to be used in Albania as well. This process will reduce the time for processing VAT file, will be used in effective way for quarterly business survey or structure business survey purposes and will increase the quality of statistical procces and product by using effectively the VAT file.