Explanatory Memorandum to COM(2011)329 - Short-term statistics as required by Council Regulation 1165/1998

Please note

This page contains a limited version of this dossier in the EU Monitor.

dossier COM(2011)329 - Short-term statistics as required by Council Regulation 1165/1998.
source COM(2011)329 EN
date 08-06-2011
52011DC0329


Contents

1.

REPORT FROM THE COMMISSION TO THE EUROPEAN PARLIAMENT AND THE COUNCIL


concerning short-term statistics as required by Council Regulation (EC) No 1165/1998 of 19 May 1998

INTRODUCTION

According to article 14 i of Council regulation 1165/1998 concerning short-term statistics:

The Commission shall, by 11 August 2008 and again every three years thereafter, submit a report to the European Parliament and the Council on the statistics compiled pursuant to this Regulation and in particular on their relevance and quality and the revision of indicators. The report shall also specifically address the cost of the statistical system and the burden on business arising from this Regulation in relation to its benefits. It shall report on best practices for lessening the burden on business and shall indicate ways of reducing the burden and costs.

The present report follows up on the report that was transmitted in June 2008 pursuant to the above article.[1] A first report on the quality of short-term statistics had already been published in January 2003.[2]

European short-term statistics provide a comprehensive set of short-term business indicators for the European economies covering industry, construction, retail trade, repair and certain service industries. The economic development of these sectors is reflected by a series of indicators for turnover, production, number of persons employed and hours worked, gross wages and salaries. In addition there are indicators for output prices, new orders and for construction costs. European short-term statistics were introduced in 1998. Since then their quality (coherence, comparability, accuracy and timeliness) and scope (indicators, countries, level of detail) have been constantly improved.

The second section of this report outlines the uses of short-term statistics and their relevance for key European policies and the steering of European monetary policy. It also indicates major developments in short-term statistics since the last quality report. The next section describes in greater detail the different quality aspects of short-term statistics. Finally, the cost and the burden imposed by the collection and processing of data for the statistics are presented along with some examples on how such cost and burden might be reduced.

A first draft of this report was presented to the members of the working group on short-term statistics in January 2011 (written consultation).

2.

Relevance of short-term statistics


The establishment of short-term statistics had been a prerequisite for the establishment of the euro area and for monitoring the European monetary policy.[3] The European Central Bank and National Central Banks are among the most important users of short-term statistics. Other important users are the Commission itself, national governments, research institutes, businesses and business organisations. News releases about short-term statistics generally receive wide attention in the media. Last but not least, short-term statistics also provide important input for other statistical areas such as national accounts.

Short-term statistics rank among the most important statistics produced by the Statistical Office of the European Union (Eurostat). Economic trend analysis, forecasting and modelling are the most common uses of short-term statistics. However, the data are also required for the preparation of policy decisions, for research purposes, to check and validate data from other sources and to prepare business decisions (e.g. in market research). In some cases results from short-term statistics are also used by businesses in rather special ways (e.g. the use of price indicators for scaling contracts).

Of the 22 Principal European Economic Indicators (PEEIs) i that have been developed to monitor the economic development of the European Union and its Member States and in particular for steering monetary policy in the euro area, nine are provided by short-term statistics, i.e. industrial production, industrial output prices of the domestic market, industrial new orders, industrial import prices, production in construction, turnover in retail trade and repair, turnover in other services, service output prices and building permits.

In the area of short-term statistics Eurostat publishes 60 news releases per year, i.e. a monthly release for five principal indicators (industrial production, industrial new orders, industrial output prices, production in construction and volume of retail trade). Moreover, each year at least five issues of 'Statistics in focus' are published which, apart from providing quantitative data, deal with selected methodological issues and analyse current economic trends. In addition, the Quarterly Panorama of European business statistics presents short-term results in greater detail.

Between 2009 and 2010 short-term statistics were the subject of a 'rolling review'. In order to identify ways to make the collection, processing and dissemination of short-term statistics more efficient and more effective, external experts analysed the framework and process for the production of short-term statistics from the point of view of users, of Eurostat and of the associated partners (national statistical institutes).

The majority of users need short-term statistics frequently and at regular intervals, and they consult the publications (database, news releases etc.) several times a month or at least once a month (e.g. when a news release is published). Accuracy is perceived by many users as the most important quality aspect of short-term statistics. However differences between the various quality aspects are not so significant and it seems fair to say that timeliness, comparability, accessibility, clarity, coherence and completeness are considered of almost the same importance as accuracy. This poses a great challenge for the producers of statistics as there are trade-offs between several quality aspects, notably between accuracy and timeliness or comparability and completeness.

In the review, users declared themselves to be satisfied with the quality of short-term statistics and the related services. 95% of the users who formulated a judgement on the quality found short-term statistics to be adequate, and 60% found it to be good or very good.

In 2008, European economies experienced the greatest downturn in economic activities since the establishment of European short-term statistics. Between April 2008 when industrial output reached its peak and April 2009 when it was at its lowest value, a reduction of almost one fifth was recorded. In the construction sector the fall was slower but also deep: output fell between January/February 2008 (pre-crisis peak) and February 2010 (lowest value) by around one fifth. The volume of retail trade also declined significantly although not as dramatically as production. These developments highlighted at the time the importance of having reliable up-to-date short-term statistics.

Two important methodological changes took place in 2009. Firstly, short-term indicators were rebased from the year 2000 to the year 2005. This involved the arithmetical change of the base year (average 100 in 2005 instead of 2000) and an update of the weights used in the indicators to reflect the economic structure of 2005. Secondly, and in parallel with the rebasing, the new classification of economic activities in the European Union, NACE Rev. 2, replaced the older version NACE Rev. 1.1. The classification was modernised by introducing new detail to reflect different forms of production and emerging industries. The detail of the classification has substantially increased in service-producing activities. NACE Rev. 2 provides a better picture of the overall economy and facilitates international comparisons. The transition to NACE Rev. 2 was accomplished without noticeable effects on quality aspects such as timeliness or stability of data (see below).

3.

Scope and quality of the short-term INDICATORS


Scope and compliance with the short-term statistics Regulation

In August 2009 the last derogations which exempted Member States from delivering data under regulation (EC) No 1165/98 to Eurostat (concerning service producer prices) expired.

Commission regulation (EC) No 329/2009 amended and updated the basic legal act and introduced hours worked and gross wages and salaries as new employment indicators for retail trade and repair and for other services. Following this amendment, hours worked and gross wages and salaries will be compiled for all branches of the economy covered by short-term statistics. Data will be calculated for quarterly reference periods and will become available in 2013.

Member States' compliance with the short-term statistics Regulation in terms of reliability, timeliness, coherence and comparability is monitored by Eurostat every six months and shows a high level of compliance and constant improvement. The EU27 average score was 8.9 (out of 10) as of 1st October 2010, as compared to 8.5 for 1st April 2005 and 6.6 for 1st January 2004. Most Member States are now close to full compliance with the Regulation.

4.

Accuracy, reliability, coherence and comparability


Regulation (EC) No 1165/98 and related acts have introduced a set of common definitions which are applied by all Member States. Eurostat and national statistical offices work together in order to ensure high accuracy, reliability and coherence of short-term indicators. The methodological framework established by the Regulation is continuously improved by mutual consultations and special thematically focussed task forces.

It should be noted that methodologies do not have to be identical in Member States. According to the short-term statistics Regulation and in accordance with the principle of subsidiarity Member States are free to decide on the most efficient and effective ways of collecting and processing data in order to take into account national differences, e.g. size, economic structure and availability of administrative data.

Eurostat also works together with other international organisations, especially the OECD, in order to increase the comparability of data and methods beyond the European Union.

5.

Timeliness and punctuality


For short-term statistics the early availability of data is of central importance. Already in 1998, the first European regulation on short-term statistics included deadlines for the delivery of national data to Eurostat (see table 1, third column). With the 2005 amendment regulation i these deadlines have been considerably shortened for most indicators. The table also indicates the current delivery dates (e.g. news release) in number of days after the end of the reference period.

In general, the timeliness of short-term statistics can be considered to be very good and any delays usually correspond to the target delivery day falling on a weekend or a public holiday. However, European releases of short-term indicators still take more time than in the United States or Japan.

Users are informed well in advance about the publication dates of the news releases by the release calendar on the Eurostat website.[6] In 2010, all dates announced in the release calendar were met.

6.

Table 1 : Timeliness measured as days between end of reference period and delivery deadline, EFC target and dissemination


Indicator Periodicity STS-Regulation deadlines a) EFC targets for 2011 b) Dissemination of EU totals c)

Industrial production Monthly

Industrial output prices of the domestic market Monthly

Industrial new orders Monthly

Industrial import prices Monthly

Production in construction Monthly

Turnover in retail trade and repair Monthly

Turnover in other services Quarterly

Service output prices Quarterly n.a.

7.

Building Permits Quarterly


a) Deadlines for the transmission of data to Eurostat after the end of the reference period according to the short-term statistics regulation. Note that for smaller countries longer deadlines often apply.

b) Targets set by the 2010 Status Report on information requirements on EMU of the Economic and Financial Committee (EFC).

c) Days between end of reference period and dissemination (e.g. news release).

8.

Revisions of short-term indicators


First results of short-term indicators are partly based on preliminary, estimated and incomplete data. It is therefore to be expected that results change between first, second and even subsequent publications. Following the first data release, survey results often become more complete because missing or late respondents have been added. The other reasons why data are revised are seasonal adjustment, benchmarking, new and/or improved data sources, and corrections of errors or methodological changes. Nevertheless, the size of revisions of short-term indicators is generally rather limited, especially at the aggregate EU level and for the euro area.

In order to assess the quality of the first results of the five short-term indicators published monthly in a special news release, the changes between the first and the second publication of monthly growth rates were analysed for the period from June 2007 until December 2010.

9.

Table 2 : Size of revision for the five principal short-term indicators a)


Indicator Average absolute revision b) Average revision c) Relative average revision d)

Industrial production 0. 0. 0.

Industrial output prices of the domestic market 0. 0. 0.

Industrial new orders 0. -0. 0.

Production in construction 0. 0. 0.

Volume of retail trade 0. 0. 0.

a) Seasonally adjusted growth rates of indicators for the euro area.

b) Average of the differences in absolute terms between growth rates at the second and at the first publication.

c) Average of the differences between growth rates at the second and the first publication.

d) Average ratio between revision and growth rate at second publication (both in absolute terms).

Table 2 shows, for the European aggregates of the five indicators subject to a news release, the average size of revisions of growth rates between the first and the second publication (one month later). In absolute terms (second column) the average revisions vary between 0.1 percentage points for industrial output prices and 0.5 points for industrial new orders. The differences between the indicators have mainly methodological reasons. While the industrial output prices are often not revised at all, revisions of the indicator for industrial production might even take place several years after the first publication.

The average revisions (third column) are close to zero. This means that, for the period analysed, there was no systematic over- or underestimation of the growth rates at the time of the first publication.

Finally, the last column in table 2 by displaying the relation between the revision of the growth rate and the growth rate itself at the time of the second publication, gives an indication of the scale of the revisions during the period analysed.

10.

Accessibility, clarity and availability of metadata


All short-term statistics results are freely accessible on the Eurostat website. The special section dedicated to short-term statistics is easy to find under the theme 'Industry, trade and services' or via the search function.[7]

Comprehensive, targeted and detailed explanations of methodological issues (metadata) are also made available, partly in the above mentioned publications (Statistics in focus, Quarterly Panorama of European business statistics) and on the Eurostat website. The database 'STS sources' provides detailed discussions of statistical processes, legal questions, confidentiality rules, data quality and description of national data collection methods.[8] For a number of key indicators additional detailed methodological explanations are also available ("PEEIs in focus").[9] Concise explanations of statistical concepts are published in a new wiki-style online dictionary and as answers to 'Frequently asked questions'.

11.

COSTS TO THE STATISTICAL SYSTEM AND BURDEN ON BUSINESSES


Eurostat conducted a general exercise covering 24 EU basic legal acts involving enterprises as respondents or as major part of the respondents in 2009 and collected data on the statistical burden on enterprises (data on production costs for statistical authorities were not included). For short-term statistics sufficient data were provided for the years 2006, 2007 and 2008 by 21 countries, but the exercise did not yield fully comparable figures because of methodological problems. Therefore only trends could be published. For short-term business statistics a relatively stable development between 2006 and 2007 was noted, followed by a slight reduction of the response burden between 2007 and 2008.

Specific cost and burden measurements for short-term statistics started with the first quality report published in 2003 which provided (for selected indicators) data on the number of enterprises answering to questionnaires as well as estimates for manpower needed by both the enterprises and the statistical authorities to produce short-term statistics. For the preparation of the second quality report published in 2008, a task force established a methodological framework to assess the cost and burden for short-term statistics as a whole and for individual indicators. The measurement tool was consistent with the EU net cost model for the measurement of the administrative burden. For the present report, data were collected between September and October 2010 for the reference year 2009. The measurement followed the methodology developed for the previous exercise but was adjusted in light of the experience gained in the past and focussed on the one hand on the time that businesses needed to compile the necessary data and to answer statistical requests and, on the other hand, on the time that statistical authorities needed to comply with the short-term statistics Regulation.

Establishing reliable and comparable quantitative indicators for the costs (for statistical systems to collect, process and disseminate data) and the burden (on enterprises to provide data) poses rather difficult practical and methodological questions. The time needed by an average business to fill in a survey questionnaire or to provide information on a particular item can only be roughly indicated. Furthermore, the number of businesses responding to individual statistical items is difficult to estimate (for instance, one business could be in several samples for different items which could lead to double counting). In accordance with the principle of subsidiarity, samples are organised differently in the Member States which makes comparisons difficult. Finally, short-term statistics cannot be fully separated from other statistics and there is often no clear method to allocate cost and burden to different statistics. Most data collected with a survey for short-term statistics will also be used for other statistics, notably national accounts. On the other hand, data collected under the nominal responsibility of other statistical departments might also be used for short-term statistics.

The available data for the European Union as a whole indicate a reduction in the production cost by 6% and a reduction in the burden on enterprises by 15% (both measured in hours per year) between 2006 and 2009. As found in the last quality report, the costs for the statistical system (e.g. national statistical offices) and the burden on businesses are roughly identical despite the slight shift of the overall cost and burden from businesses towards statistical authorities.

A tentative calculation of the burden (stemming mainly from filling in questionnaires for short-term statistics) on an average business is presented in table 3 for the principal indicators. Note that for the ease of comparison the burden is calculated as minutes per month even for the quarterly indicators. A minimum burden of zero indicates that data are not collected with a short-term statistics questionnaire but are taken from administrative sources. There are still several cases where the burden appears to be overestimated.

12.

Table 3: Unweighted country-averages (EU 27), highest and lowest country values of the burden (minutes per month) on a business replying to a questionnaire


Industrial production Industrial output prices Industrial new orders Production in construction Turnover in retail & repair

13.

Average


Maximum

Minimum

IMPROVEMENTS IN THE COLLECTION AND PROCESSING OF SHORT-TERM STATISTICS

There is a bundle of factors that explain why the cost reduction and the rough data available do not allow a quantitative distinction of the various influences on the total figure. In several cases, sample sizes for surveys have been reduced or postal or fax surveys have been replaced by electronic and internet surveys. This leads to a reduction of the costs for the statistical authorities. The positive effect for businesses stems from the possibility to include plausibility checks in the electronic questionnaire which results in fewer mistakes and a lower number of queries regarding classifications.

A major reason for the reduction in the burden on businesses appears to be the increased reliance by statistical offices on administrative data, especially for the employment indicators (hours worked, persons employed, gross wages and salaries). This kind of information is often already available e.g. within tax offices, labour administrations, public social insurances or building authorities and does not have to be collected again for statistical purposes. While the use of existing administrative sources for statistical purposes reduces the burden on businesses it may lead to higher costs within statistical offices as data might have to be adjusted. The use of administrative data also requires continuous coordination especially where countries have decentralised administrations.

Moreover, the use of administrative data as input for short-term statistics is limited especially for the monthly indicators where the requirements in terms of timeliness and short deadlines are very high. Administrative data are often collected less frequently than would be necessary for short-term statistics and/or become available too late. In addition, adjustments are required where administrative data do not directly fit the definitional requirements of short-term statistics, and such adjustments can also have a bearing on data quality.

Short-term statistics are continually improved by the statistical offices on national level and in cooperation between national offices and Eurostat in several task forces. Together with some other statistical domains short-term statistics are covered by the programme for the Modernisation of European Enterprise and Trade Statistics (MEETS). Within this programme short-term statistics are included in several actions concerning the further use of administrative data, the accommodation to sources not yet useable, the combination of data from different sources, the data-warehousing, the estimation of non-available data, and the harmonisation of definitions. These processes of identification, development and sharing of already existing good practices should contribute to an increase in efficiency of data collection and processing.

Short-term statistics already use a variety of instruments to limit the burden on enterprises. For some indicators European sample schemes have been introduced. Countries participating in such a scheme only have to compile data for those industries or products for which their national contributions to the aggregate at European level make a significant difference.

Another important instrument to reduce or even avoid burden on small businesses is applying thresholds when collecting data. In cases where small enterprises cannot be completely exempted from the data collection, lower sampling rates are applied, i.e. for a small business it is less likely to be included in a sample. Another sampling technique that is routinely applied by many statistical offices is the rotation of samples so that a share of the businesses is replaced every year. It should also be pointed out that sample sizes required for European short-term statistics are often lower than the sample sizes used in practice, since national offices may have to reproduce the indicators with a regional or local breakdown.

Annex: The 43 short-term indicators (PEEI in bold print), Q – quarterly indicator, M – monthly indicator

14.

Industry Construction Retail trade and repair Other services


Production Industrial production (M) Production in construction (M/Q) Building construction Civil engineering - -

Turnover Industrial turnover (M) Domestic industrial turnover Non-domestic industrial (Euro area and Non Euro area) - Turnover in retail trade and repair (M) Turnover in other services (Q)

Costs - Construction costs (Q) Material costs Labour costs - -

New orders / building permits New orders received (M) Domestic new orders Non domestic new orders (Euro area and Non Euro area) Building permits (Q) number of dwellings m2 of useful floor area - -

Number of persons employed Number of persons employed in industry (Q) Number of persons employed in construction (Q) Number of persons employed in retail trade and repair (Q) (not yet available) Number of persons employed in other services (Q) (not yet available)

Hours worked Hours worked in industry (Q) Hours worked in construction (Q) Hours worked in retail trade and repair (Q) (not yet available) Hours worked in other services (Q) (not yet available)

Gross wages and salaries Gross wages and salaries in industry (Q) Gross wages and salaries in construction (Q) Gross wages and salaries in retail trade and repair (Q) Gross wages and salaries in other services (Q)

Output prices Industrial output prices (M) Domestic output prices Non domestic output prices (Euro area and Non Euro area Deflator of sales (M) Output prices of other services (Q)

Import prices Industrial import prices (M) (Euro area and Non Euro area) - - -
href="http://circa.europa.eu/Public/irc/dsis/ebt/library?l=/methodology/sts_sources&vm=detailed&sb=Title" target="_blank">circa.europa.eu/Public/irc/dsis/ebt/library?l=
href="http://circa.europa.eu/Public/irc/dsis/ebt/library?l=/methodology&vm=detailed&sb=Title" target="_blank">circa.europa.eu/Public/irc/dsis/ebt/library?l=