Macro Price setting in the euro area: Some stylised facts from Individual Producer Price
ABSTRACT This paper analyzes the relationship between banksâ€™ divergent strategies toward specialization and diversification of financial activities and their ability to withstand a banking sector crash. We first generate market-based measures of banksâ€™ systemic risk exposures using extreme value analysis. Systemic banking risk is measured as the tail beta, which equals the probability of a sharp decline in a bankâ€™s stock price conditional on a crash in a banking index. Subsequently, the impact of (the correlation between) interest income and the components of non-interest income on this risk measure is assessed. The heterogeneity in extreme bank risk is attributed to differences in the scope of non-traditional banking activities: non-interest generating activities increase banksâ€™ tail beta. In addition, smaller banks and better-capitalized banks are better able to withstand extremely adverse conditions. These relationships are stronger during turbulent times compared to normal economic conditions. Overall, diversifying financial activities under one umbrella institution does not improve banking system stability, which may explain why financial conglomerates trade at a discount.
- SourceAvailable from: Luis J Alvarez[Show abstract] [Hide abstract]
ABSTRACT: This paper shows that the standard Calvo model clearly fails to account for the distribution of price durations found in micro data. We propose a novel price setting model that fully captures heterogeneity in individual pricing behavior. Specifically, we assume that there is a continuum of firms that set prices according to a Calvo mechanism, each of them with a possibly different price adjustment parameter. The model is estimated by maximum likelihood and closely matches individual consumer and producer price data. Incorporating estimated price setting rules into a standard DSGE model shows that fully accounting for pricing heterogeneity is crucial to understanding inflation and output dynamics. The standard calibration that assumes within sector homogeneity, as in Carvalho (2006), is at odds with micro data evidence and leads to a substantial distortion of estimates of the real impact of monetary policy.The B.E. Journal of Macroeconomics. 01/2010; 10(1):13-13.
Article: Indirect ICT investment[Show abstract] [Hide abstract]
ABSTRACT: Numerous economic studies have highlighted that information and communication technology (ICT) investment expenditure appears to be greater in the field of services, and above all in certain activities such as financial services and wholesale trade, than in industry. This analysis examines whether the investment data compiled by national accountants underestimate productive ICT expenditure, and whether this is more pronounced in the case of industry than for services. For that purpose, we propose an assessment of the size of ICT inputs (termed 'indirect ICT investment') in non-ICT investment expenditure for France in 2000. The main result of our analysis is that the amount of 'indirect ICT investment' appears to be small, compared with 'direct ICT investment', suggesting that the biggest investors in ICT remain services, regardless of whether we consider 'direct' or 'direct' plus 'indirect investment'.Applied Economics Letters 11/2009; 16(17):1713-1716. · 0.23 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Our survey covers the recent developments of the microeconometric literature on evaluation methods. In this field, the canonical model is Rubin's causal model, which is close to Roy's selectivity model. This model is the relevant framework for defining and for examining the identifiability conditions of the parameters of interest in any evaluation study. We insist on the definition of these parameters, which include the average effect of the treatment on the treated and on the non-treated individuals. For each set of assumptions (selectivity on observable or unobservable characteristics, conditional independence between outcomes and treatment indicators, etc.), we present the most adapted estimation method. We put a special emphasis on matching estimators in the situation where the selectivity depends only on observables, and on differences-in-differences methods and on regression-discontinuity techniques when the selectivity depends both on observable and unobservable characteristics.Banque de France, Documents de Travail. 01/2007;
ET DE RECHERCHE
DIRECTION GÉNÉRALE DES ÉTUDES ET DES RELATIONS INTERNATIONALES
PRICE SETTING IN THE EURO AREA:
SOME STYLISED FACTS FROM
INDIVIDUAL PRODUCER PRICE DATA
Philip Vermeulen, Daniel Dias, Maarten Dossche, Erwan Gautier, Ignacio
Hernando, Roberto Sabbatini, Harald Stahl
NER - E # 164
DIRECTION GÉNÉRALE DES ÉTUDES ET DES RELATIONS INTERNATIONALES
DIRECTION DE LA RECHERCHE
PRICE SETTING IN THE EURO AREA:
SOME STYLISED FACTS FROM
INDIVIDUAL PRODUCER PRICE DATA
Philip Vermeulen, Daniel Dias, Maarten Dossche, Erwan Gautier, Ignacio
Hernando, Roberto Sabbatini, Harald Stahl
NER - E # 164
Les Notes d'Études et de Recherche reflètent les idées personnelles de leurs auteurs et n'expriment pas
nécessairement la position de la Banque de France. Ce document est disponible sur le site internet de la
Banque de France « www.banque-france.fr ».
Working Papers reflect the opinions of the authors and do not necessarily express the views of the Banque
de France. This document is available on the Banque de France Website “www.banque-france.fr”.
Price setting in the euro area: Some stylised facts from
Individual Producer Price Data1
Philip Vermeulen, Daniel Dias, Maarten Dossche, Erwan Gautier,
Ignacio Hernando, Roberto Sabbatini, Harald Stahl2
1 The views expressed in this paper are those of the authors and do not necessarily reflect the views of neither the ECB nor the National
Central Bank to which they are affiliated.
This paper is based on a set of studies, each related to a euro area country, conducted in the context of a Eurosystem research project
(Inflation Persistence Network, hereafter IPN). The authors belong to the ECB and to the National Central Banks that have been involved in
a research sub-group of the IPN devoted to the analysis of micro producer prices. The contribution of many other members of this research
group and co-authors of country studies, (Luis Álvarez, Pablo Burriel, David Cornille, Monica Dias, Silvia Fabiani, Angela Gatulli, Claire
Loupias, Pedro Neves, Patrick Sevestre and Giovanni Veronese), without whom this study would not have been possible, is strongly
acknowledged. The authors would also like to thank the national statistical institutes for providing the data, and the many members of the
IPN and an anonymous referee for helpful comments.
2 Corresponding author: email@example.com
Authors affiliations: European Central Bank (Philip Vermeulen), Banco de Portugal (Daniel Dias), National Bank of Belgium (Maarten
Dossche), Banque de France (Erwan Gautier), Banco de España (Ignacio Hernando), Banca d’Italia (Roberto Sabbatini), Deutsche
Bundesbank (Harald Stahl)
Correspondant Banque de France : Erwan Gautier (firstname.lastname@example.org)
Cet article résume un ensemble d’études sur la fixation des prix à la production menées dans 6 pays
de la zone euro: Allemagne, France, Italie, Espagne, Belgique et Portugal. Il rassemble les résultats
obtenus dans chacune des études nationales et de nouveaux résultats pour la zone euro. Ces études
utilisent des données de relevés de prix mensuels. Cinq faits stylisés apparaissent communs à
l’ensemble des pays. Les prix à la production changent assez peu fréquemment : chaque mois environ
21% des prix sont modifiés. La fréquence des changements de prix est très hétérogène entre les
secteurs : les prix changent très fréquemment dans l’énergie, moins souvent dans les secteurs de
l’alimentaire et les biens intermédiaires et assez rarement dans les secteurs des biens durables et des
autres biens finis. L’hétérogénéité entre les pays est plus faible et les différences dans le degré de
flexibilité entre les secteurs sont communes à l’ensemble des pays. Il n’y a pas de rigidité à la baisse
des prix à la production : 45% des changements de prix sont des baisses et 55% des hausses.
L’ampleur des changements de prix est assez importante comparée au taux d’inflation. Cet article
examine aussi les différents facteurs pouvant expliquer les changements de prix. La structure des
coûts, la concurrence, la saisonnalité, l’inflation et les prix « psychologiques » jouent tous un rôle
dans la décision de changement de prix. Enfin, les prix à la production semblent plus flexibles que les
prix à la consommation.
Codes JEL : E31, D40, C25
Mots-clé : Fixation des prix, prix à la production
This paper documents producer price setting in 6 countries of the euro area: Germany, France, Italy,
Spain, Belgium and Portugal. It collects evidence from available studies on each of those countries
and also provides new evidence. These studies use monthly producer price data. The following five
stylised facts emerge consistently across countries. First, producer prices change infrequently: each
month around 21% of prices change. Second, there is substantial cross-sector heterogeneity in the
frequency of price changes: prices change very often in the energy sector, less often in food and
intermediate goods and least often in non-durable non- food and durable goods. Third, countries have
a similar ranking of industries in terms of frequency of price changes. Fourth, there is no evidence of
downward nominal rigidity: price changes are for about 45% decreases and 55% increases. Fifth,
price changes are sizeable compared to the inflation rate. The paper also examines the factors driving
producer price changes. It finds that costs structure, competition, seasonality, inflation and attractive
pricing all play a role in driving producer price changes. In addition producer prices tend to be more
flexible than consumer prices.
JEL Codes: E31, D40, C25
Keywords: Price-setting, producer prices
Cet article résume un ensemble d’études sur la fixation des prix à la production menées dans
6 pays de la zone euro: Allemagne, France, Italie, Espagne, Belgique et Portugal. Il rassemble
les résultats obtenus dans chacune des études nationales de l’ « Inflation Persistence
Network » (un réseau de recherche de l’Eurosystème) et de nouveaux résultats pour la zone
euro. Les données microéconomiques de relevés de prix à la production utilisées pour
construire les indices de prix à la production ont été mobilisées dans ces études.
Les cinq faits stylisés suivants apparaissent communs à l’ensemble des pays considérés :
1. Les prix à la production changent assez peu fréquemment: chaque mois environ 21%
des prix sont modifiés.
2. Il existe une importante hétérogénéité dans la fréquence des changements de prix
entre les secteurs : les prix changent très fréquemment dans l’énergie, un peu moins
souvent dans les secteurs de l’alimentaire et les biens intermédiaires et encore moins
souvent dans les secteurs des biens durables et des autres biens finis.
3. L’hétérogénéité entre les pays est plus faible et les différences dans le degré de
flexibilité entre les secteurs sont communes à l’ensemble des pays.
4. Il n’y a pas de rigidité à la baisse des prix à la production : 45% des changements de
prix sont des baisses et 55% des hausses.
5. L’ampleur des changements de prix est assez importante comparée au taux
Cet article examine aussi les différents facteurs pouvant expliquer les changements de prix
comme la structure des coûts, la concurrence, la saisonnalité, l’inflation et les prix
« psychologiques ». Nous montrons ainsi qu’une part du coût du travail dans les coûts de
production plus importante conduit à des fréquences de changements de prix plus faibles. Au
contraire, une part plus importante des consommations intermédiaires hors énergie et énergie
correspond à des fréquences de changements de prix plus importantes. Nous obtenons aussi
qu’une concurrence plus élevée est associée à des prix plus flexibles. Le taux d’inflation est
positivement corrélé avec les fréquences de changements de prix. Nous trouvons enfin que
les changements de prix sont plus fréquents en janvier et qu’ils sont moins nombreux en août
Enfin, nous effectuons une comparaison de la rigidité des prix à la production et des prix à la
consommation. Nous trouvons que les prix à la production semblent plus flexibles que les
prix à la consommation.
Ces nouveaux résultats permettent de mieux comprendre la fixation des prix à la production
et doivent aider la modélisation macroéconomique à mieux comprendre les effets de la
In this paper, we document producer price setting in 6 countries of the euro area: Germany,
France, Italy, Spain, Belgium and Portugal. We collect evidence from available studies of the
Inflation Persistence Network (a network of researchers of the Eurosystem) on each of those
countries and also provide new evidence. These studies use monthly producer price data from
price records used to construct official producer price indexes.
The following five stylised facts emerge consistently across countries:
1. Producer prices change infrequently: each month around 21% of prices change.
2. There is substantial cross-sector heterogeneity in the frequency of price changes:
prices change very often in the energy sector, less often in food and intermediate
goods and least often in non-durable non- food and durable goods.
3. Countries have a similar ranking of industries in terms of frequency of price changes.
4. There is no evidence of downward nominal rigidity: price changes are for about 45%
decreases and 55% increases.
5. Price changes are sizeable compared to the inflation rate.
We also examine the factors driving producer price changes such as costs structure,
competition, seasonality, inflation and attractive pricing. In particular, we find that a higher
share of labour costs in the total cost of manufacturing the product corresponds to a lower
frequency of price changes where on the contrary, a higher share of non-energy intermediate
inputs and of energy goods correspond to a higher frequency of price adjustments. We also
find that a higher degree of competition is associated with more flexible prices and that
higher inflation is positively correlated with a higher frequency of price adjustment. We also
show that price changes occur more often in January. By contrast, price changes tend to occur
less over the summer months, particularly in August, and in December. In addition we
compare producer prices with consumer prices.
We find that producer price tend to be more flexible than consumer prices. This indicates that
the retail level adds an additional level of stickiness to prices above the producer level.
The new results in this paper broaden our understanding of producer price setting and should
help macro economic modelling and ultimately be beneficial for conducting monetary policy.
During the period from 2003 to 2005, the Eurosystem embarked on a wide study of
inflation persistence and price stickiness in the euro area (Inflation Persistence Network,
IPN). The IPN analysed price setting practices in the euro area by looking at various
databases never exploited in previous empirical research. In particular, micro consumer
prices were analysed (Dhyne et al., 2006) as well as information on price setting was
collected through firms’ surveys (Fabiani et al., 2006). This paper presents a set of new
empirical insights in producer price behaviour based on the analysis of detailed micro data
provided for the first time by National Statistical Institutes. It brings together results obtained
in national studies and it produces empirical evidence for the euro area as a whole, based on a
coordinated approach. The approach and the structure of the paper is closely related to the
complementary analysis of consumer prices by Dhyne et al. (2006); similarly, its main scope
is to collect the stylized facts on producer price setting that can be derived from the analysis
of the available data.
The emphasis of the paper is on the rigidity of prices. How rigid are producer prices? Do
they change often or not? The existent literature on the nature of consumer price setting
(Cecchetti, 1986; Kashyap 1995) for a limited set of goods has recently been revived by new
evidence stemming from much broader datasets spanning the U.S. and Euro area CPI (Bils
and Klenow, 2004; Dhyne et al., 2006). The general finding of those studies is that consumer
prices are relatively rigid. Whereas consumer prices are of course relevant for the monitoring
of inflation by central banks, it is the prices at the producer level that are ultimately modelled
in economists macro-economic models. For instance, rigidity of prices measured by the
frequency of price setting is a key element of new Keynesian models that traditionally
describe producers as Calvo price-setters (Yun, 1996). A deeper knowledge of producer price
setting should help macro economic modelling and ultimately be beneficial for conducting
This paper describes the characteristics of producer price setting behaviour based on
price records used to construct official producer price indexes. Where they are relevant,
complementary findings from one-time surveys of producers are mentioned. In the first part,
the behaviour of individual product prices at the establishment level is investigated. The
empirical assessment of the periodicity and size of individual price changes using micro price
data has been scarce and partial until very recently, due to the limited amount of data
available to researchers. Earlier micro-studies on price setting referred mostly to consumer
prices and focused on a very limited number of products. The evidence based on individual
producer prices is even scarcer. Stigler and Kindahl (1970) and Carlton (1986), analysing
transaction prices of intermediate products used in manufacturing, are among the few micro-
studies on producer prices. Carlton’s (1986) findings indicate quite rigid producer prices. A
larger literature studies the managerial decision making processes and practices that are
involved with price changes of producers. For instance, the work by Bergen et al. (2003), and
Zbaracki et al. (2004) shows that changing prices is a quite involved process that often
includes costly information gathering, decision making, communication and customer costs,
at least for large enterprises. This could explain the observed rigidity.
More recently, in the context of the IPN, some researchers have exploited the large-scale
data sets of individual prices underlying the official Producer Price Index (PPI). In particular,
the statistical offices of Germany, France, Italy, Spain, Belgium and Portugal allowed
researchers to investigate individual price records under strict confidentiality agreements. In
this paper their findings on the behaviour of individual producer prices are collected and
The paper is structured as follows. Section 2 presents the main characteristics of the
databases. Section 3 provides a set of stylized facts on producer prices that can be derived
from the data. Section 4 analyses the determinants of producer price changes. Section 5
compares the flexibility of producer prices with consumer prices. Section 6 concludes.
2. MICRO QUANTITATIVE PRODUCER PRICES FOR THE EURO AREA
The statistical offices of individual countries collect monthly price records on
individual products at the establishment level to construct the producer price index at the
industry and country level. The monthly collection by the statistical offices of price records
of products sold by all domestic establishments is done by means of a statistical survey, that
is, price records are obtained from a representative sample of establishments and products.
The national indices constructed on the basis of the individual price data are further
aggregated to obtain euro area wide producer price indices. In Europe, the price record data
collection is harmonized by a Directive (that is a European law) of the European Union. In
particular, the methodological manual from Eurostat (Eurostat, 2002) explains that the
following rules apply for the collection of prices by the statistical offices:
• The appropriate price is the ex-factory price including all duties and taxes except
value added tax (VAT).
• All price-determining characteristics of the products are taken into account, including
quantity of units sold, transport provided, rebates, service conditions, guarantee
conditions and destination. The specification of the product must be such that in
subsequent reference periods, the establishment is able uniquely to identify the
product and to provide the appropriate price per unit of the product.
• The prices are actual transaction prices, not list prices.
• The price collected in period t should refer to orders booked during period t not the
moment when the commodities leave the factory gate.
• If transport costs are included, this should be part of the product specification.
All statistical offices apply these rules so that price records are comparable across
countries. Notwithstanding the above, we are aware that statistical offices are likely not to be
able to follow strictly the guidelines for all products at all times (e.g. a list price might be
used if no transaction occurred during the month); consequently, there might be some random
variation left due to procedures internal to statistical offices, we are not aware of. There
might also be some random variation (and even errors) in the reporting by establishments.
Note for instance that the guidelines from Eurostat for the price record taking do not say
anything about whether the establishment has to follow the same customer over time (if
possible). Some establishments with long term relationships may report prices for the same
product and the same customer, month after month (so that prices might not change much),
whereas other establishments may have varying customers month after month.
Although it is a priori possible that part of the differences across countries in the
statistics provided in this paper could be due to methodological rather than economic
differences, we do not believe this to be a major issue. In addition, as it will be shown below,
the fact that statistics are remarkably similar across countries is reassuring of the possibility
of deriving broad stylized facts that are relatively robust for the euro area as a whole. In all
countries, researchers were able to follow the price of a product at a particular establishment.
Price records in all countries contained at least the following information: the actual price, a
product code, an establishment code, a code indicating product replacement, and the year and
the month of the record. By following prices for a given product from the same
establishment, price trajectories are observed. The product code for Germany, France, Italy
Belgium and Portugal is the PRODCOM code, which is the official classification code of
products produced within the European Union, whereas for Spain it is a numeric sub-variety
code which prevents identification of the specific product for the researcher.
Monthly quantitative price records, namely individual price trajectories, that is sequences
of price quotes for a specific product from a specific establishment, were made available for
Germany, France, Spain, Italy, Belgium and Portugal, all together accounting for a weight of
around 87% of the euro area PPI. Researchers in these countries had access to nearly the
complete set of micro data underlying the computation of the national PPI, with the exception
of Italy where only a representative subset of price records referred to 60 products was made
available. A complete and detailed description of each national database is provided in
country analyses (Table 1).
Table 1 - Coverage of the national databases
Country Paper Percentage of PPI basket
covered in the national
Belgium Cornille and Dossche (2006) 83 January 2001- January 2005
France Gautier (2006) 92 January 1994- June 2005
Germany Stahl (2006) 100 January 1997 - February 2003
Italy Sabbatini et al. (2005) 441 January 1997- December 2002
Portugal Dias, Dias and Neves (2004) Almost 100 January 1995 – December 2000
Spain Álvarez et al. (2005) 99.4 November 1991 - February 1999
(1) Estimated on the basis of 3-digit weights (see Sabbatini et al., 2005)
For all countries, each individual price record corresponds to a precisely defined
product, manufactured by a particular establishment in a particular month and year. The
products included in the PPI basket can be classified in 6 different product categories: food
products, non-durable non-food products, durable products, intermediate goods, energy and
capital goods. Appendix A contains the classification of NACE-3 digit industries into those 6
Before analysing the characteristics of the price setting behaviour in the euro area, it is
useful to recall in which inflation environment this study takes place. The average yearly
inflation, as measured by the aggregate producer price index over the period of the respective
databases was 1.0% in Germany, 0.7% in France, 1.5% in Italy, 2.1% in Spain, 1.5% in
Belgium and 1.7% in Portugal. Hence in all countries this was a low inflation period.
However it has to be kept in mind that average inflation hides the fact that PPI inflation is
generally quite volatile from month to month.
3. THE CROSS-SECTIONAL AND TIME SERIES PATTERNS OF PRODUCER
This section presents a set of stylised facts on the cross-sectional and time series patterns
of producer price changes in the euro area. The main statistic used is the monthly frequency
of price changes, whose magnitude is compared across countries and industries. The monthly
frequency of price changes can be defined as the share of prices that are changed in a given
month. Say 100 establishments provide the prices of bricks of clay in month t-1 and month t.
If 20 of the prices differ from t to t-1, the frequency of price changes of bricks of clay in
month t is 0.20. Clearly, the frequency of price changes can be calculated at different levels
of aggregation across good categories (for individual products such as “bricks of clay”, for
items belonging to the same category such as “NACE 264, Manufacture of bricks tiles and
construction products, in baked clay”, for higher aggregate categories such as NACE 26
“Manufacture of other non-metallic mineral products”, in the extreme case for all
manufactured goods) and across time periods.
A specific problem arises when calculating frequencies at different levels of aggregation:
frequencies at higher levels of aggregation are derived by weighting those calculated at lower
levels of aggregation. For instance, the frequency of price changes at aggregate NACE 26 is
calculated as a weighted average of the frequencies of the subgroups of NACE 26, that is
NACE 261, NACE 262, and so on up to NACE 268. Furthermore, the frequency of each of
those subgroups, say NACE 264, is a weighted average of the frequencies of the products
belonging to the subgroup NACE 264. At the lowest level of aggregation usually no weights
are available, so that all products in that subgroup get the same weight. All statistics in this
paper are calculated using country specific PPI weights. Country PPI weights differ as
countries do not have the same industrial structure; that is some products or product
categories are produced more in some countries than in other countries. In Appendix B we
present more formally how the frequency of price changes has been computed.
The frequency of price changes of a particular product or product group provides
condensed information on the outcome of price setting. Clearly it has to be interpreted with
caution, since the frequency may not be independent of the causes of price changes. If a
particular product has a very low frequency of price changes this could be due to the fact that
it is not flexible at all (that is it does not react promptly to causes) or that it does not need to
be adjusted since the underlying factors driving the price level do not change.
Fact 1 – Producer prices change rather infrequently. The frequency of monthly price
changes ranges from 0.15 in Italy to 0.25 in France.
Table 2 provides the (average weighted) frequency of price changes for all goods.
We find an average frequency of price changes for the euro area of 21%, higher than the
average frequency of price changes of 15% found by Dhyne et al. (2006) for consumer
prices. The reference to average frequencies is, however, not a reliable indication of the
differences in the degree of price stickiness, as the composition of the CPI and PPI baskets
differs considerably. A detailed comparison of frequency differences between consumer and
producer goods is given in section 5.
The frequency of Germany, France, Spain, Belgium and Portugal are all lying in a
narrow interval between 0.21 and 0.25. The highest frequency occurs in France (0.25), the
lowest in Italy (0.15). However, for Italy energy products are excluded, whereas they usually
have the highest frequency of price changes; this narrows the above range, implying that the
average weighted frequency across euro area countries is very similar. However, the lower
frequency of Italy cannot be fully explained by the absence of energy products. In fact, as we
discuss later, the frequency of price changes for different product categories tends to be
smaller in Italy than in the Euro area (table 3). Moreover, when looking at 19 different 2 digit
NACE industries, 16 have a higher frequency in Germany than in Italy, so that it appears that
producer prices in Italy are somewhat stickier.
Table 2: Frequency of price changes all goods
Belgium 0.24 0.13 0.11
0.25 0.14 0.11
0.22 0.12 0.10
0.15 0.09 0.07
Portugal 0.23 0.14 0.10
Spain 0.21 0.12 0.09
Euro area2 0.21 0.12 0.10
(1) Energy prices are excluded. – (2) The euro area is calculated
using the relative weights of total industry producer price index
of the euro area (domestic).
Fact 2 – There is a substantial degree of heterogeneity in the frequency of price changes
across industries, which can be classified in three broad classes. Price changes are very
frequent for energy products, relatively frequent for food and intermediate products and
relatively infrequent for capital goods, non-durable non-food and durable products.
Table 3 shows the frequency of price changes according to 6 product categories. From
this table it is clear that the frequency of price changes is heterogeneous across and within
main industrial groupings and across countries. Energy prices change most frequently in all
countries, which is due to oil products in the energy component. The euro area frequency of
price changes for energy is 72%. Food prices, with a euro area frequency of 27% as well as
intermediate goods, with a euro area frequency of 22 %, also change quite often. On the
contrary, capital goods prices (euro area frequency of 9%), non-durable non- food (euro area
frequency of 11%) and durable goods prices (euro area frequency of 10%) change least
frequently. The fact that energy prices change most frequently is likely due to volatile supply.
When looked into more detail it seems that there are frequent price changes for products that
are simple and have not undergone a series of transformations. This is consistent with the
observation in Bils and Klenow (2004) that prices of raw goods are changed more often than
processed goods. This implies that the costs of those products are closely linked to the
corresponding raw material price which is presumable set daily on exchanges. A case is, for
example, the frequency of price changes of “flour” and “bread”: such frequencies are above
40% and equal to 6%, respectively, both in Italy and in Portugal. Other products with
generally high frequency of price changes are, for instance, textile fibers, paper and paper
board, veneer sheets, plywood, dairy products, non-ferrous metals, metal wires, sugar, coffee,
etc. All these products have undergone little transformation from input to end product. The
heterogeneity of price changes across industries seems therefore akin to the heterogeneity
across products and product groups found by Dhyne et al. (2006). For instance, in the CPI,
energy prices, and unprocessed food have the highest frequency, two categories of products
that have undergone little transformation. On the other hand, capital goods, non-durable non-
food and durable products generally consist of a whole series of inputs such as raw materials,
labour, R&D, etc.
Table 3 – Frequency of price change by product category (1)
Non- durable Durable
non- food products
Belgium 0.20 0.11 0.14 0.28 0.50 0.13
0.32 0.10 0.13 0.23 0.66 0.12
0.26 0.14 0.10 0.23 0.94 0.10
Italy 0.27 0.10 0.07 0.18 na 0.05
Portugal 0.21 0.05 0.18 0.12 0.66 na
Spain 0.24 0.10 0.10 0.28 0.38 0.08
Euro area 0.27 0.11 0.10 0.22 0.72 0.09
(1) For each component, the euro area figure is computed as the average of the national results, weighted with
the national weights of the considered sub-index in the euro area PPI.
The average weighted frequency of price changes masks a lot of heterogeneity across product
groups. To document this heterogeneity, the frequency of price changes was calculated at the
2-digit industry level for each country according to the NACE classification (from NACE 15
to NACE 36). The distribution of those 2-digit industry level frequencies, for all countries
jointly (unweighted), is represented in Figure 1.3 The mode of the distribution is around 0.09.
The distribution also shows large outliers of high frequencies. The highest frequencies of
price change correspond to “Manufacture of refined petroleum products” (NACE 23)
(frequencies above 85%) and “Manufacture of basic metals” (NACE 27) (frequencies above
The distribution is wide for all countries. This is illustrated in Table 4 which reports
country specific minimum and maximum two digit industry level frequencies of price
changes as well as the country specific standard deviation of the distribution of the two-digit