Rounding errors and index numbers
The purpose of this article is to show how
index numbers are being systematically distorted
by significant rounding errors in the revised
or rebased data. After a brief discussion
of why index numbers are especially vulnerable
to rounding errors, the CPI for urban
wage earners and clerical workers (CPI-W) is
used to illustrate how large these errors can be
and how they now distort the pattern of
monthly changes in the 1960s. Next, a Monte
Carlo study is presented to show some of the
econometric problems associated with rounding
errors of this magnitude. Finally, I recommend
steps to better understand and/or alleviate
this problem.