Discussion of the paper “Testing Time Series Data Compatibility for Benchmarking”, by Benoit Quenneville and Christian Gagné

Benchmarking is an important component of the reliability of macroeconomic information, and one which is relevant for both the final data user and the econometrician, as it deals with combining information from different sources, characterised by different frequencies of observation and different degrees of reliability. The paper by Quenneville and Gagné proposes tests for checking the compatibility of the series as a step preliminary to benchmarking. This discussion focuses on the actual implementation of these tests, on alternative tests available in the econometric literature, and on the various sources of measurement errors that affect benchmarking.