Mr David Mair

Collection of prices of
recreational goods and electronic
and electrical appliances in the
EU
David Mair, Head of Unit on Consumer
Markets, DG SANCO
for the Conference
Statistical architecture for consumer prices
Luxembourg
15-16 October 2009
Contents
1.
2.
3.
4.
5.
The objectives of the study
Contractor and his offer
Description of the source data
Problems encountered and
solutions taken
Deficiencies of data and collection
The objectives of the study
Collection of prices of goods that are:
• Comparable and substituable from a
consumer perspective (not necessarily
identical models)
• Representative: product (may be
represented by different substituable
models) shall be for sale in the
overwhelming majority of the Member
States and have significant market
share within the product group
Contractor and his offer
GfK Marketing Services Ltd., London
Monthly or bi-monthly syndicated collection of
prices by models in shops
Collection covers max 21 countries (Austria,
Belgium, Bulgaria, Czech Republic, Denmark,
Finland, France, Germany, Greece, Hungary,
Ireland, Italy, Netherlands, Poland, Portugal,
Romania, Slovakia, Slovenia, Spain, Sweden,
United Kingdom)
For some categories and some countries data not
collected or confidential (if 1 or 2 sellers only)
10 product groups covered by
the offer
Recreational goods and household electronics
• Flat screen TVs
• Digital cameras
• Mp3
• Notebooks
• Refrigerators
• Filter Coffee makers
• Irons
• Microwave ovens
• Vacuum cleaners
• Washing machines
Description of the source data
• Syndicated collection – number of
shops:
• for consumer electronics from 247 in
Slovenia to 10.584 in Germany
• for household appliances from 390 in
Slovenia to 12.606 in Italy
• for mobile computers from 550 in Ireland to
7510 in Germany
• Other shops: extrapolation from the
sample – problem of sample choice
• Prices calculated as sales/units sold
(»natural average » of prices)
Description of the source
collection of prices cont.
• Sales channels cuts available: Consumer
electronics shops/Mass Merchants;
online/offline
• No regional structure
• Monthly/bimonthly collection (depending
on type of product)
• Problem of seasonality not solved
• Problem of confidentiality in particular
cases (e.g. exclusive distributors)
Problems encountered
• Period to choose – annual chosen,
to avoid seasonality bias
• Number of models (from 2.641 for
digital cameras to 21.951 for
mobile computers) – scattered
nature of the market
• Comparability vs. representativity:
very few models actually sold
across Europe
Solution to comparability vs
representativity
• 2 levels of comparability – models and
« technical segments » (groups of
models of the same features)
• 3 levels of representativity (sold in all
countries, in most, in big majority)
• The « most adequate » (sufficiently
covering the market) selection chosen as
principal
Problem of structural differences
• Situation: in any aggregate (average
price per category, per group of models)
strong impact of structures
• Solution: weighted structures (by
European average structure, by the
structure of the complementing channel)
or fixed prices (« technicity index »)
• Problem: « holes » (Swiss cheese) –
models of groups of models missing in a
country of sales channel
Deficiencies of the collection –
quality of data
• Sample of points of sales not randomly drawn.
Not possible to estimate margins of errors
• Some countries not covered
• Problem of seasonality not solved
• Problem of calculation of prices out of sales and
units sold
• Impact of structure of sales
• Impact of “bundles” (e.g. computers for 1
euro)
• Products returned –sales may be negative
Deficiencies of the collection –
coverage and power relations
• Only 2 providers of datasets –
Nielsen, GfK
• monopolist position vs. distributed
clients
• impact on price of service
• clients should be collective (Eurostat?)
• Only some sectors covered
• Nielsen – groceries
• GfK - electronics
DG SANCO consumer website
http://ec.europa.eu/dgs/health_consumer/consumer_en.htm
Thank you