Abstract
A consumer price index number (CPI) is the result of combining data from various complex sample surveys. In the present paper we concentrate on but one aspect of the accuracy of a CPI, namely, the variance due to the sampling variability of the budget survey(s) from which the weighting coefficients are derived. Using the individual budget survey data the variance of a CPI can be estimated by means of the method of balanced half samples. This method is applied to several types of CPIs and to CPIs for different groups of households.
Get full access to this article
View all access options for this article.