The Purchasing Power Parities (PPPs) for international comparisons are computed in benchmark years: for example the last two International Comparison Program (ICP) data collections took place between 1993 and 1996 and in 2005. The PPPs for non-benchmark years are estimated by extrapolating the PPPs from the most recent ICP round according to the deflators of GDP and its components in each country considered.

When the extrapolated PPPs are compared with those based on the last ICP round, high differences between them are usually found. As many authors pointed out, this issue is particularly important since the estimates of GDP (at PPP) must be revised for many countries, and therefore a great deal of debate over the consistency between the two sets of results arises.

Although the PPPs from two subsequent benchmarks are not completely comparable due to the many changes in data collection procedures and methodology used for their computation, the high differences in the results cast doubt on the comparability of the resulting PPPs over time and ask for the interpretation of differences both from a statistical and an economic point of view.

The aim of this paper is twofold. Firstly, to discuss the methods used and proposed for the temporal extrapolation of the PPPs. Considering in particular the household consumption, an improved method for extrapolating the PPPs, using the data collected for the computation of Consumer Price Indices (CPIs) at Basic Heading level, is suggested. In this context, a spatial-temporal statistical model could improve the PPP extrapolation. Secondly, in order to explain, both statistically and economically, the divergences between the extrapolated PPPs and computed PPPs in the last benchmark a methodological approach is proposed. This approach is based on the decomposition of the formulae into components which refer to the prices and to the system of weights respectively.

Keywords: Extrapolation of PPPs; Use of CPIs data; Divergenses between the extrapolated PPPs and the computed PPPs; Method of decomposition

Biography: Professor Luigi Biggeri is graduated in Economics with specialization in Statistics. and obtained a MA in Management. He is full professor of Economic Statistics since 1980. He is giving lectures in two course, Business Statistics and Statistical Methods for the evaluation of policies and programs.

He has been president of the Italian Statistical Society, of the Italian Statistical Institute (Istat) and of the International Association of Survey Statisticians (IASS).

He published more tha 170 scientific works in price index numbers, purchaising power parities, productivity and efficiency of public administration units, indicators for the evaluation of universities, organisation of surveys and of the National Sstatistical Offices.