This week two new papers were published by UNIDO STAT staff:

The UNIDO ISID working paper "Value added in industrial statistics and national accounts. Equal in theory, different in practice" by Martin Haitzmann and Valentin Todorov addresses the frequently asked question about the deviations in published value added figures. UNIDO compiles and provides value added estimates on total manufacturing in its databases and publications, which are generally derived from two different sources. On the one hand, detailed structural business statistics data in industrial statistics (commonly based on surveys and censuses) provide figures on the value added of manufacturing industries (INDSTAT databases). On the other hand, manufacturing value added measures an exclusive and exhaustive contribution of manufacturing to the gross domestic product of an economy and therefore refers explicitly to national accounts concepts (MVA database). While it is correctly assumed that these two figures need to be equal in theory, they differ in practice, even if structural business statistics data are among the most important data sources in many countries and play a central role in national accounts.

The second paper, "Monitoring Robust Estimates for Compositional Data" by Valentin Todorov, published in the Austrian Journal of Statistics (online first), considers the problem of exploratory data analysis and estimation in the case of compositional data. Compositional data sets are characterized by multivariate observations (vectors) containing relative contributions of parts to a whole. Examples are geochemical composition of rocks, household budget patterns, time budget, ceramic compositions, but also value added, output or any other indicator measured over the manufacturing industries in INDSTAT or the technological structure of exports. To illustrate the problem of monitoring compositional data and to demonstrate the usefulness of monitoring in this case a real-life data set presenting the technological structure of manufactured exports as it is used in the UNIDO CIP index is analysed.


[1] Haitzmann M. and Todorov V. (2020) Value added in industrial statistics and national accounts: Equal in theory but different in practice. UNIDO ISID Working Paper, Vienna, pdf.

[2] Todorov V. (2021)  Monitoring Robust Estimates for Compositional Data. Austrian Journal of Statistics, 50, pp. 16-37, pdf.


Query name already exist
Download Data
The requested download exceeds the max allowed rows for Excel 2010 - please narrow you selection and try again.
Download Data
Requested Data expired. Please, request new download.