The estimation of substitution bias in the CPI

Estimation of Consumer Price Index (CPI) and approximation of this estimation to true Cost of Living Index are very important issues due to fact that CPI is a critical indicator for formation of an economic policy. Employment of fixed based period indexes, like Laspeyres, by statistical agencies cause to bias in indexes. These biases namely are, commodity substitution bias, outlet substitution bias, new goods substitution bias and quality change bias. Such kind of measurement difficulties would lead to bias in true cost of living index. A probable bias in CPI should not be underestimated considering wide and functional usage of this index.In Turkey there has not been an extensive study on biases in indexes like this thesis. Also a bias value in CPI has not been estimated and published in Turkey. The purpose of this study is to estimate the bias in CPI in Turkey. The literature part of this study gives general information on indexes and estimation methods of CPI and their applications in Turkey and in other countries. Moreover, the study explains bias types in indexes, estimation methods of the biases and ways of reducing of these biases. In the application part of the thesis, 2003=100 based CPI is used as a data source. The bias values are estimated on yearly bases in general CPI and in indexes of main expenditure groups for the 2004-2009 periods. An approach based on subtraction of Fisher index, a superlative type, from Laspeyres index is employed for the bias estimations.The findings show very small biases for general CPI on yearly bases. On the other hand the biases in main consumption groups like Alcoholic beverages and tobacco products, Housing and Education are found to be sizeable amount. These biases are seen to be related with sudden price changes in the items used for estimation of main expenditure group indexes. The sudden price changes in these items are shown with graphics so as to support this conclusion.Key Words: CPI, Cost of living index, Substitution bias, Outlet substitution bias, New goods bias, Quality change bias