Research Work Group on CPI measurement. Higher School of Economics, Perm branch,
Perm, Russia.
The
aim of the group is to get a deeper
understanding of the Russian CPI methodology,
evaluate potential errors and biases.
The work of the group
was organized in the following way: first of
all, a lot of time was devoted to a review of
articles and BLS, StatCanada, Goskomstat manuals
on CPI. Than we analyzed the methods and results
of relevant for us articles. And finally
analogous research was done: this stage also
included data preparation. All that process was
accompanied with a two-times-in-a-month
presentation.
The first presentation was made at November
9, 2005.
Main Presentations,
prepared by me. Unfortunately it occurred to be
very difficult to collect all the material and
presentations due to the break of a computer,
where all material were
located:
9
November2005 "CPI: methodology, errors and
biases, method of elimination".
View
27
January 2006: "Social CPI in Russia: weights".
View
1
February 2006 "Social
CPI in Russia: Methodology".
View
February 2006
"Hedonic Indices
for PC in Russia".
View
April
2006 "Methodology of elementary indices in
Russia".
View
11
May 2006: "Methodology
of elementary indices in Russia-2".
View
5
June 2006 "Research
Agenda for the next year".
View
August
2006 "Final
Presentation, summarizing main results with the
projections for the next year".
View
Calculations,
a core essential part of our activities. At the
first stage of our research project we spent a
plenty of time trying to understand more deeply
Russian CPI (Goskomstat) methodology. To do that
we've calculated social (minimum) CPI with usage
of Goskomstat high-level (aggregated) data for
food. This step allowed us to identify potential
errors and biases which can arise both at high
and low level of CPI calculation and data
preparation. Our results have shown no
significant difference between our and official
numbers. All calculation were done for the perm
region to avoid to intra-region aggregation -
our task were to find as
disaggregate data as possible. But we do
understand it could lead to a bigger standard
deviation of an estimate:
Winter 2005 "CPI:
Perm region, Goskomstat Data".
View
Then we wanted to find out how such problems
as new good/ quality change, new outlets which
usually lead to a commodity and outlet rotation can
bias CPI. This analysis were performed for the
relativity homogenous (eggs, milk and cheese) and
heterogeneous commodities (Personal Computers). A
significant expenditure weight is a main reason why
we've chosen the first set. Personal Computers are
implemented in the Russian CPI calculation several
years, but you can hardly find a price index for PC.
Also, a lot of literature suggest that there's an
upward bias for PC price index. Both facts in
combination with the growing wiliness of Russian
Government, President and Central Bank to reduce
inflation means an increasing demand for more
precise estimates.
Winter/Spring 2005/06. Homogenous Products
"Data
and intermediate calculations".
View
"Final
CPI calculations".
View
"Comparison
of our estimates
with official ones".
View
Winter/Spring 2005/06. Heterogeneous Products
(PC)
"Price
and characteristics data for 2005".
View
"Price
and characteristics data for 2004".
View
"Hedonic
Price Indices
Estimates for 2005".
View
"Bachelor
of Science in Economics Dissertation - "Hedonic
Indices for PC in Russia".
View
"Bachelor
of Science in Economics Dissertation - "Hedonic
Indices for PC in Russia".
View
Manual
on CPI were
written by me for the students willing to join
the Working Group on CPI as a way of
facilitation - brief outlook combined with
examples on CPI calculation gives a complex
picture of CPI methodology, its major stages and
difficulties which can arises at these stages:
"Manual
on Russian CPI: a Brief Guide for Students".
View
