52008DC0340

[pic] | COMMISSION OF THE EUROPEAN COMMUNITIES |

Brussels, 9.6.2008

COM(2008) 340 final

REPORT FROM THE CO MMISSION TO THE COUNCIL AND THE EUROPEAN PARLIAMENT

concerning Short-Term Statistics as required by Regulation (EC) No 1165/98 of the European Parliament and of the Council of 19 May 1998

REPORT FROM THE COMMISSION TO THE COUNCIL AND THE EUROPEAN PARLIAMENT

concerning Short-Term Statistics as required by Regulation (EC) No 1165/98 of the European Parliament and of the Council of 19 May 1998

Regulation 1158/2005 of the European Parliament and of the Council concerning short-term statistics amending the Regulation 1165/98 introduces in article 14 that "t he 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 of business and shall indicate ways to reduce the burden and costs".

The main conclusions of this report can be summarised as follows: the first section recalls that the statistics produced pursuant to this regulation are the key short-term economic statistics indispensable for the conduct of European economic and monetary policy. Their availability, coverage, comparability and timeliness have increased substantially as a result of the regulation; the ambitious targets set in the EMU action plan in 2002 have been achieved. The second and third sections show that the quality of the indicators is generally good for the purposes for which they are intended and revisions are kept within reasonable bounds. Finally, the last section demonstrates that the cost of these statistics and the burden on enterprises are modest and describes best practices that have been implemented to minimise the burden. To comply with Article 14, the Commission (Eurostat) set up a Task Force of national statistical offices with the mandate to develop a tool to measure the burden on businesses and the cost to the statistical system arising from the STS Regulation. This tool was developed in consistency with the EU net cost model and in collaboration with the Member States and referred to the Commission commitment to reducing burden[1].

In 2007, the task force drafted a text assessing the benefits of the STS Regulation, covering several items related to the EU and euro area economy: the increasing amount of statistical information, the timeliness, the scope of statistics, the quality of statistics, the revision of statistics, the punctuality of statistics, the possibilities to compare it with e.g. the US and Japan. The text was submitted to the STS main users at European level and country level in spring 2007.

The draft report was discussed and agreed in broad terms with the experts in the Working Group on Short-Term Statistics and in the Statistical Programme Committee.

Section 1: Relevance of short-term statistics and the benefits arising from the STS Regulation

The European short-term statistics Regulation (STS-R) (adopted in 1998 and substantially amended in 2005[2]) introduced a common set of indicators and definitions for the EU Member States in order to better analyse the business cycle at EU and national level, partly as a precondition for the establishment of the euro area with its common monetary policy. The statistical requirements for the formulation and monitoring of the euro area monetary policy were expressed by the ECB and the Commission on many occasions[3]. The European Central Bank (ECB), the National Central Banks as well as the Commission and national governments in the EU have high user needs as regards short-term statistics. The successful implementation of the STS-Regulation has been a milestone for the provision of high-quality short-term economic indicators for the euro area as a whole. The structured requirements of the STS-R allowed Eurostat systematically to produce European aggregates.

In November 2002, the Commission submitted to the European Parliament and the Council a comprehensive report on euro area statistics supporting the development of the Principal European Economic Indicators (PEEIs) and their full implementation. The PEEIs cover a list of nineteen infra-annual macro economic indicators for the euro area and the EU, for which challenging improvement objectives were set. STS provides eight[4] out of nineteen of the PEEIs (see table 1). Not only the seven available Principal European Economic Indicators (PEEIs) but all the 39[5] indicators monitored through the STS Regulation provide on a timely basis an accurate and summarized picture of the EU and euro area economy.

The STS indicators constitute 40% of all regular News Releases by Eurostat and are systematically among the most consulted on the web site. Eurostat produces 60 monthly News Releases per year on short-term indicators (industrial production index, industrial output price index, industrial new orders index, production in construction, turnover index for retail trade).

Increasing amount of statistical information on the EU and euro area economy

The improved data supply has allowed a significant increase in the use of STS for economic analysis. Whereas in 1999, the launch of Stage Three of EMU, industrial production was the only regularly available monthly European STS indicator, in 2007 intensive use is made of a range of monthly and quarterly indicators on output, demand, prices, wages and employment, for industry, construction, retail trade and market services. The data are used for current economic analysis at EU and euro area level, for country analysis and country comparisons, for forecasting, economic research and for economic and monetary policy decisions. For economic analyses and policy users the STS data provides a unique dataset, supplementing other statistical source like quarterly national accounts or qualitative (opinion) surveys. Finally the STS dataset has been recently extended to the services sector. It helps to meet an important policy need as the importance of services in the economy continues to increase but the benefits of this implementation will be fully operational only in 2009 due to derogations introduced with the 2005 amending Regulation (see footnote 4).

The volume of data also increased considerably due to the recently acceding Member States and the transmission of more data from the EU15 Member States. With data transmitted by 31 countries (27 Member States, Turkey and Croatia, Norway and Switzerland), the number of series is now four times bigger than 5 years ago.

T imeliness of statistics on the EU and euro area economy

To satisfy greater needs for timely data and analysis of the European Monetary Union, the 1998 STS Regulation introduced deadlines for transmission. The 2005 amending Regulation went further by introducing significantly shorter deadlines. As Table 1 shows, the new deadlines are almost systematically in line with the "EU target deadlines" set by the Communication of the Commission to the European Council and Parliament on euro area statistics (COM(2002)661). The timeliness in STS improved in the last years and has been achieved with systematic and gradual development work in the Member States in order to maintain a high level of data quality. However some EU indicators' releases are still behind the US and Japan dissemination dates.

The 2005 amending Regulation introduced some European sampling, which is particularly suitable for first and timely estimates at European level and, therefore, fulfil major requirements of users while minimising the cost and burden in each Member State.

Table 1 : timeliness in days after the reference period of the Principal European Economic Indicators related to STS

List of PEEIs for business indicators | Periodicity | in 2002 | Target for 2008 | current delivery[6] (as of May 2007) |

3.1. Industrial production index | Monthly | 48 | 40 | 40 |

3.2. Industrial output price index for domestic markets | Monthly | 35 | 35 | 33 |

3.3. Industrial new orders index | Monthly | Did not exist | 40/50 | 50 |

3.4. Industrial import price index | Monthly | Did not exist | 45 | 45 (planned for 2007) |

3.5. Production in construction | Monthly | 75 and quarterly | 45 | 45 (since Jan 2007) |

3.6. Turnover index for retail trade and repair | Monthly | 60 | 30 | 30 first estimate / 60 details |

3.7. Turnover index for “Other services” | Quarterly | partial | 60 | 60 |

3.8. Output prices index for services | Quarterly | Did not exist | 60(*) | 90 partial data in 2008 |

(*) The original target was superseded by the 2005 Regulation which stipulates the mandatory deadline for transmission of the output prices for services at 90 days.

Scope and quality of statistics on the EU and euro area economy

European data are available earlier. At the same time the country coverage at the moment of their first release has improved. This is an important quality feature as it reduces the risk of subsequent revisions of first releases. Many users judged the actual timeliness good and recommended not to go for further acceleration to the detriment of the accuracy: stable first estimates provide more benefits to users than quicker but less stable results.

The MEETS program[7] may help Member States and the Commission (Eurostat) identifying new needs concerning the scope of the STS Regulation.

The quality of short term statistics has been assessed by users as generally good or satisfactory. The short-term statistics are assessed as unbiased and accurate.

R evision of statistics on the EU and euro area economy

Revisions of STS have been generally limited, in particular at euro area and EU level; however, for some of the indicators (e.g. retail trade turnover) further work is needed in order to improve the reliability of first estimates.

National and European users assessed the revisions to be of appropriate size, underlining that the size may not be the most important factor compared to the frequency of revisions and should not be the outcome of speeding up collection and estimation processes.

P unctuality of statistics on the EU and euro area economy

Like most EU countries, Eurostat has established an advance release calendar for many of the indicators, which is important for users of the data. The release dates pre-announced on the Eurostat website for the whole year are respected. The punctuality of STS statistics is unanimously assessed as optimal, very good or satisfactory. The respect of advance release calendars is very important regarding the integrity of statistical practices.

C omparison between the EU and euro area economy and e.g. the US and Japan.

The Council Regulation, the ensuing Commission Regulation on definitions and the methodological guidelines are all applied by the countries transmitting STS data. This regulatory and methodological framework (including common tools like classifications, harmonised adjustment techniques to remove calendar and seasonal effects) ensures a good comparability between national data as well as good-quality European aggregates.

However compilation processes and methodological choices are not compelled to be 100% equal from one country to another in order to allow countries to make best use of the data available to them. In the most cost-effective way, countries may apply different collection methods (surveys, use of administrative sources) and different calculation modes for the data.

The correspondence in terms of definitions, methods and quality between the EU and euro area economies and other major economies is very good and permits better economic analysis. Eurostat works closely with OECD and other international institutions to ensure comparable methodologies and data and the use of common tools.

Comparing the timeliness of some statistics, the US and Japan remain faster than European data (employment, labour force and wages were mentioned). However, the right balance between quality and timeliness should be ensured.

Section 2: Quality of the data

Five principles of the European statistics code of practice, adopted in 2005, are devoted to the quality of statistical output: relevance, accuracy and reliability, timeliness and punctuality[8], coherence and comparability[9], accessibility and clarity.

Through the consultation of national and European users during spring 2007 the relevance of short-term statistics data has been assessed as good by the users: the short-term statistics meet their needs (see section 1). Most of the STS indicators are available for more than 10 years at European level, allowing business cycle analysis. The compliance of Member States in terms of reliability, timeliness, coherence and comparability is monitored by Eurostat every six months and shows a continuous improvement: the EU27 average score was 9.0 (out of 10) as of 1st April 2007, 8.5 (1st April 2005) and 6.6 (1st January 2004). Most Member States are close to full compliance with the STS Regulation.

The accuracy of the statistical output is tackled at national and Community levels, by eliminating as much as possible non-sampling errors (bias due to definitions, processing errors, measurement errors, coverage errors), by calculating sampling errors and studying and analysing revisions (see section 3 on revisions).. More detailed technical information is given in publications for the Industrial Production Index, the retail trade turnover index and the domestic output price index.

All the short-term statistics data are accessible on the Eurostat website. Since 2006 a dedicated section on the Eurostat website relates to short-term business statistics[10] providing the latest publications, access to data, background information (legal and methodological) and answers to frequently asked questions. Every month short-term statistics news releases are among the top 10 downloads and hits of the Eurostat website. Clarity is ensured by a comprehensive and coherent set of metadata, tables and graphs related to published series. The metadata is publicly accessible. Since 2000, Eurostat created a database names "STS sources & methods" gathering meta information from all Member States and all indicators. The STS sources and methods (last updated in 2007) provide information on coverage, periodicity, timeliness, publication, legal basis and confidentiality rules, quality, and methodology.

Section 3: revisions

Revisions of STS have been generally limited, in particular at euro area and EU level; however, for some of the indicators further work is needed in order to improve the reliability of first estimates. A group of experts from OECD, ECB, Eurostat and Member States is working within a Task Force to unify the revision policies between indicators and countries.

There are different sources of revisions in short-term statistics. STS indices are based predominantly on survey data. Therefore one source of revision at national level is the late responses to surveys from some companies. First estimates could be derived from judgmental or statistical techniques and may have to be revised when more data become available or as a result of benchmarking.

A second source of revision is methodological changes (changes in statistical methods, changes in concepts, definitions or classifications) which can take many forms: the data sources have been improved (a new survey or administrative data have been developed), some existing practices coming from European or international requirements have been implemented.

A third source of revision is the statistical adjustments done at national level and at European level to take account of the seasonality and the calendar effects affecting time series.

Finally, revisions could occasionally be due to accidents such as errors in source data or computation processes.

Table 2 gives the sizes of revision in growth rates between the first published index and the second release (one month later) for European aggregates of the five indicators subject to news releases. Among those five indicators, the industrial new orders and the construction production have the largest revision measured by the average over two years period of the absolute revision. The industrial domestic output price index has the smallest revision.

Table 2 : Size of Revision for five short-term indicators - EU level

Mean absolute revision[11] | Average later growth rate | Mean revision | Relative mean absolute revision[12] |

Industrial production SA[13] | 0.1 | 0.3 | 0.0 | 0.18 |

Industrial new orders SA | 0.4 | 0.8 | 0.2 | 0.23 |

Industrial domestic output prices | 0.1 | 0.3 | 0.0 | 0.27 |

Construction production SA[14] | 0.2 | 0.4 | -0.2 | 0.43 |

Retail trade deflated turnover SA | 0.1 | 0.2 | 0.0 | 0.27 |

The revision policy may depend on the source. For example, in some countries the labour input indicator (number of persons employed) comes from the Labour Force Survey which is a household based survey and the estimates are generally not revised.

In the industrial sector, the output price index is often not revised. The production index is an indicator with the largest gap between the first published estimate and the final data: from 3 months to 2 or 3 years because of benchmarks with structural or National accounts data.

In the retail trade and services sector the employment indicator is more subject to a long revision period than turnover. In some countries short-term employment indicators based on business survey data are finalised after 14 months to 2-3 years after the reference quarter for consistency with structural data.

Hence, revision policies vary from one short-term indicator to another and from country to country. However the implementation of STS Regulation and especially its amendment in 2005, by setting up demanding deadlines for transmission from the Member States, has driven the Member States to strengthen the whole process of compiling short-term statistics and to minimise the revisions.

Section 4: Cost and burden; best practices to lessen the burden on businesses

Cost and burden

In 2005 a task force was set up among short-term statistics specialists to develop a tool to measure the statistical burden[15] on businesses and the cost to the statistical system arising from the STS Regulation according to a common European methodology. The tool was developed in consistency with the EU net cost model. However the measures are only expressed in hours worked per year and not translated into euros.

The measure of burden is defined as the time the business needs to compile all the data to answer the statistical request. The measure of cost is defined as the time needed at the competent national statistical authority for all activities to comply with the STS Regulation.

The measures have been finalised during the first quarter of 2007. Based on twenty-six countries, a good overall picture for the European Union is obtained.

In a typical year, around 930,000 European businesses responded to a statistical survey related to the 39 short-term statistics indicators. On average each business responding spent 4 hours 20 minutes per year (21min per month) to fill in the statistical surveys related to the STS Regulation and the 39 indicators. Respondents are mainly larger enterprises because small and medium sized businesses are generally exempted (see section 4.2). Overall (for the 39 indicators over one year of data collection and calculation) the cost for the statistical system is about the same size as the burden on businesses. Because of the predominance of the industry sector (21 indicators among the 39) in the STS Regulation itself and in the data collection system, the number of businesses involved in the compilation of industrial data represents 44% of the total number (respectively 38% in the retail trade and services sector and 18% in the construction sector).

Monthly indicators contribute a significant share (80%) to the total statistical cost and burden but provide at the same time high benefits for the users: monthly indicators are released to ensure timely economic analysis.

It should be stressed that the costs described above are the total costs of compiling these statistics, not the extra costs of the existence of the EU Regulation. The STS Regulation came into force in 1998 at a time when already many short-term statistics were implemented in the EU countries. Moreover more than two-thirds of the short-term indicators covered by the STS Regulation are input for other EU Regulations (National Accounts, Labour Cost Index) or output from other EU Regulations (Labour Force Survey).

Best practices to lessen the burden

In fulfilling the STS requirements, Member States strive to minimise the statistical burden on business while producing data of sufficient timeliness and reliability for economic analysis. Practices and procedures have been implemented to this end in many countries and are being developed in others. Some examples are given below.

Some 20 countries introduced electronic data collection as an option offered to businesses instead of the post or fax for returning the questionnaires. Electronic data collection gives flexibility to the companies for answering within certain delays with a first data validation at the company's level. Some National Statistical Offices (NSOs) implemented an interface with standard software for company accounts widely used by businesses.

Administrative data sources are part of input used by the NSOs to compile the STS indicators. The obvious advantage is that there is no additional statistical burden imposed on companies, even though the administrative burden remains.

The VAT or other fiscal or administrative data are a very broadly-based source of information on the business system. Their use for statistical purposes implies multiple advantages, in particular the reduction of the statistical burden on companies Conversely there can be some loss of data quality if the concepts are too different from statistical needs or change over time. There have been several cases where an administrative source suddenly disappeared leaving the statisticians with no option but to recreate a statistical survey.

Where administrative data are used to compile the STS indicators, the VAT register or declarations account for roughly 50% and an employment register for the other half. Administrative data sources are more often used for quarterly data since they are generally not as timely as purpose-designed monthly statistical surveys.

Frequently, administrative data sources are used to supplement the survey data, providing information relating to small or medium-sized businesses.

All Member States avoid burden on small businesses by including thresholds or implementing sample strategies. Regarding indicators under the STS Regulation, the countries apply thresholds in the data collection for many indicators – a vast majority of small businesses are for this reason excluded from the sample, which is drawn anyhow to be representative of the total population and ensure a very good coverage of the variable of interest[16]. Low sampling rates are applied mainly in retail activities for small business (below 10 persons employed) and may result in less accuracy of the final estimates. To reduce the burden on small businesses some estimation procedures may be used such as forecast of a monthly pattern from a quarterly inquiry, or using information from administrative sources (such as VAT).

Technical sampling solutions like rotation samples (for example, 20% of the businesses are replaced every year) are applied quite often by larger Member States. The sample sizes are optimised using advanced statistical allocation procedures. Sample strategies have to be put in relation to the demand: sometimes on top of a national indicator required by the EU Regulation may be added regional or local indicators to satisfy national user needs, necessitating much larger samples.

The coordination of surveys among the STS indicators as well as between them and other surveys is given an important role inside the National Statistical Offices to reduce the overlap of statistics and so the burden on companies by not asking the same question several times.

Some National Statistical Offices have experimented successfully with a "one entry data point" especially for big companies: the large business profiler programme. This holistic strategy creates very good long-term relationships between businesses and statistical offices, rationalizes the data requirements of the statistical office and creates conditions to manage the response burden.

The STS Regulation itself permits a substantial reduction of the burden by introducing the concept of European Sample Schemes for some variables. For these cases, countries may opt to compile data exclusively for those industries or products that are very important and make a significant contribution to the aggregates at European level. The 2005 amending Regulation introduced as well thresholds for small Member States . Both of these techniques reduce the burden on enterprises while still allowing the Commission (Eurostat) to calculate reliable EU aggregates.

Other European initiatives include methodological guidelines or specific workshops to create synergy among the European statistical community and promote the exchange of best practices and discussions. In the framework of the 2005 amending Regulation, Eurostat provided, in coordination with OECD, methodological guidelines for the implementation of services output prices indices and organized several workshops on specific services activities.

Annex: list of short-term statistics indicators (PEEIs in bold)

Sector | Indicator | Periodicity M: Monthly Q: Quarterly | Regulation deadline[17] |

Industry | Production | M | 1 month 10 days |

turnover, domestic turnover, non domestic turnover (split euro area – non euro area) | M | 2 months |

new orders received, domestic new orders , non domestic new orders (split euro area – non euro area) | M | 1 month 20 days |

number of persons employed | Q | 2 months |

hours worked | Q | 3 months |

gross wages and salaries | Q | 3 months |

output prices, domestic output prices, non domestic output prices (split euro area – non euro area) | M | 1 month 5 days |

import prices (split euro area – non euro area) | M | 1 month 15 days |

Construction | Production, production of building construction, production of civil engineering | M / Q[18] |

number of persons employed | Q | 2 months |

hours worked | Q | 3 months |

gross wages and salaries | Q | 3 months |

construction costs, material costs,labour costs | Q | 3 months |

building permits: number of dwellings; building permits: m² useful floor area | Q | 3 months |

Retail trade | turnover | M | 1 month[19] / 2 months |

number of persons employed | Q | 2 months |

deflated turnover | M | 1 month / 2 months |

Other services | Turnover | Q | 2 months |

number of persons employed | Q | 2 months |

output prices | Q | 3 months |

[1] COM(2005) 97 "Better Regulation for Growth and Jobs in the European Union"

[2] Regulation (EC) n° 1158/2005 of the European Parliament and of the Council of 6 July 2005 amending Council Regulation (EC) n° 1165/98 concerning short-term statistics.

[3] For example, “Statistical requirements of the European Central Bank in the field of General Economic Statistics”, ECB, August 2000.

[4] Services output prices indicators are not yet available at the European level because of lengthy derogation requested by most Member States up to August 2008.

[5] List in annex.

[6] The 2005 amended Regulation granted an extra 15 days delay for smaller Member States.

[7] The Modernisation of European Enterprise and Trade Statistics (MEETS) program is a six-year program (2008-2013). Its first objective is to develop target sets of indicators and to review priorities.

[8] Cf. 1.2 and 1.5

[9] Cf. 1.6

[10] Under the theme Industry, trade and services

[11] The revision is the difference between the first published index and second estimate one month later. The data covers the period from November 2005 to May 2007. For example, the growth rate of industrial production for the month of February 2007 for ¬ area compared to January 2007 was estimated at 0.6 in the first release. One month later, the value of the growth rate was estimateay 2007. For example, the growth rate of industrial production for the month of February 2007 for €area compared to January 2007 was estimated at 0.6 in the first release. One month later, the value of the growth rate was estimated to 0.5, corresponding to a revision of -0.1 and an absolute revision of 0.1 to the growth rate.

[12] The relative mean absolute revision is the ratio between the sum of the absolute revision and the sum of the absolute later estimates.

[13] SA stands for seasonally adjusted; Gross corresponds to gross data.

[14] The construction production index is monthly since November 2006. Revisions are calculated on quarterly data from 1st quarter 2005 to 3rd quarter 2006 and then on monthly data from November 2006 to June 2007.

[15] Only the burden on businesses coming from statistical surveys is evaluated. When administrative sources are used to compile short-term statistics, the burden was not considered in this exercise as it is incurred anyway for administrative purpose.

[16] In the industrial sector, thresholds for exclusion of small business from the survey may be applied (less than 10 persons employed very often) but the coverage is maintained at a very high level (above 70-80%). When activities are less concentrated like in construction, retail trade or restaurants, thresholds of total exclusion are not applied. The burden on small business is lessened by implementing different sampling rates according to turnover or employment data.

[17] The deadline may be up to 15 calendar days longer for the smaller Member State.

[18] Depends on the size of the Member State.

[19] One month for aggregates, two months for details.