Demography Of Small and Medium Enterprises

1st Working Group Meeting

Report of Meeting

1. Opening and welcome

Peter Struijs, as chairman, introduced John Perry as the project manager.

2. Agenda

The agenda, document 2, was agreed

3. Contents and planning of the project

This session was covered by two documents: 3 being the agreed project proposal; and document 4, being a timetable for the new project.

The project proposal had been circulated to members of the PECO Panel project. The proposals had now been agreed and the first part had, in principle, been allocated Phare funds as part of a wider Eurostat programme of pilot projects. The actual funding had not yet been agreed by Brussels. Although not a new experience with projects, this has caused problems with starting the project. These had been solved in part by shifting funds from the previous, PECO Panel, project. It was the intentions to carry on with this approach to ensure continuity between the projects.

A separate contract for the B3 survey had been drafted to permit continuity of central data processing by Infostat. Further central data processing would need to go to tender.

A definition of “SMEs” was needed for the project. This could not be the Commission definition, because the threshold of that definition was considered too high. Anyway, the defnition of SMEs would only apply to the output of the project; for the selection of units for surveys no threshold would apply, among others because the size of units is one of the variables to be measured.

On other aspects of the project, further delays in funding will affect the finalisation of contracts with participating countries. The possibility of these delays could not be excluded. Existing funding will continue to be used as far as possible to allow the working group to continue with the preparations for the B3 survey. The situation would need further review if delays in funding continued.

The second part of the project had not yet been agreed. It was presented as an annex to the first part so that it was clear that there was a strong link with the (agreed) first part. It would be necessary to make a formal submission towards the end of 1998. The chances of acceptance were high.

Poland stated that their statistical law and constitution require early decision on surveys to be included in next year’s programme. There was also a serious problem in submitting the B3 for this year’s programme. Any change required the signature of the Prime Minister. This was likely given that the survey was sponsored by Eurostat. Hungary has similar problems, in that the survey needs to be accepted into the data collection programme. Hungary pointed out that the directors of the national statistical institutes had actively supported the project in a meeting in Vilnius towards the end of 1997. This too would help. Both countries thought that the problems could be solved but an early decision (before the Summer) would be valuable.

Bulgaria raised questions in detail about the timetable. In particular, they wished to know about the role of the seminar on used of business registers in helping with estimation and imputation in the survey. They also felt that the start date for the B3 survey was too early.

It was agreed that the project required an acronym: DOSME was suggested.

4. Organisational arrangements for the project and communication

In the event there was no document 5 for this item. The key features were continuity with the PECO Panel project in selection of Infostat for the B3 processing. A new feature was the appointment of an external (to Eurostat) project manager.

A number of important communication issues raised on 22 April we re-examined. Communications should be by Email. It was important to ensure that people received the relevant documents. It was important that Infostat and the project manager met between working group meetings. Infostat would in some instances communicate directly with countries without involving the project manager. A web site could be helpful in documenting and disseminating documents but not for transferring data files to individual countries.

The project manager would test Email links with a contact in each country by sending a message with some attached MS Word 6 files. For future meetings of the working group papers would be sent using the same method five working days before the meeting.

On the question of a web site, confidentiality issues needed to be addressed. All countries stated that they had access to the Internet. One possibility was to have a web site containing an index to documents. All documents could be held as MS Word 6 files linked to this index and protected by passwords. Management of a site was an important issue. This work required resources and could not be done prior to the next working group meeting. Infostat would examine the issues.

As far as software was concerned for Survey B3 processing:

	Existing Infostat programs would be used for data entry
	Although there was the possibility of moving from Infostat programs to commercial software (SOLAS), their were costs and licensing issues associated with a change. In addition, a change might adversely affect the plans to move imputation from Infostat to the individual countries
	For imputation countries also needed access to commercial statistical analysis software (preferably SAS or SPSS).
	Database management was an issue affecting only Infostat. They could investigate the possibility of moving to a commercial relational database management program.
	PECO Help documentation currently used software purchased originally by Phillippe Pommier but now operated by Infostat. Infostat would look at the possibility of moving from this to HTML or some other system but would initially simply update the existing help for use with Survey B3.

6. Survey B3

Document 6 contained proposals for sample sizes. Given the proximity of the start date, it was agreed that FYROM would not form part of the B3 sample. Instead a small scale pilot would be conducted in parallel with Survey B3.

The questionnaire was discussed with the following conclusions:

	No questions would be added
	Although there was some doubt about the usefulness of 26 (regarding bookkeeping), no change would be made to this question. It was important, however, not to impute for non-response to this question. There was also a need to analyse the results of this question from B3 and earlier surveys to inform Survey B4 and later surveys.
	The word “additional” would be inserted before “job” in question 32 to improve clarity.
	Slovenia, who had experienced particular partial non-response problems would report to the project manager by 4 May on possible issues affecting the questionnaire design.

The timetable in document 4 would be amended in the following ways:

	References to CEECs would be changed to CECs
	A new item would be added before 1 to cover the provision of sampling programs to CECs by Infostat. Infostat would require one week to change the reference periods and to re-compile and test. The date for this event would be 1/5/98.
	Event 1 would change to event 2 and later numbers changed accordingly. Infostat suggested that the date be changed to 15/5/98.
	Event 2 would change as a consequence to 22/5/98
	Event 17 would change to 1/10/98 to allow CECs more time to prepare the survey following the summer holiday period. The survey reference date of 1/9/98 would not change.
	Event 18 would change to refer to sending “files” to Infostat and dates would also change as a consequence of the change to event 17.
	Event 30 would change to place CECs first to reflect the fact that they would be doing the imputation. The word “partial” would be dropped. The date would be set at 28/2/99.
	Events would be inserted prior to the 2nd working group meeting to cover the supply of imputation programs to CECs by Infostat and the supply of documentation to CECs by Paul Smith.
	Event 37 would mention only English as the publication language to reflect the relative lack of interest in French and German versions.

7. Contractual matters

Document 7 formed the basis of this item. This showed a draft of the proposed contract. This would apply to eleven countries. A separate contract would be set up with FYROM. A total value for each contract would be determined prior to 22 May. The value would be based on a price per unit of 8 ECU; this would be explicitly mentioned in the contracts. As in previous surveys, this payment would be made if a direct or indirect response of 97% was achieved.

The contract includes work on imputation by the countries. The additional work on the preparation of population estimates was not covered in the contract as it would be done as part of the 2nd working group meeting.

There was some concern about delays in payment by CESD that were not remedied by article 6. Peter Struijs felt that the contract could not be strengthened by a late payment clause but would speak to CESD to ensure prompt payment. Receipt notifications for invoices could perhaps be introduced. It was possible that some delays had been the result of incomplete information appearing on invoices. The contract would be amended to included a statement regarding the precise requirements for invoices. Payments would be made only after approval by the project manager and verification by Eurostat D1.

Contracts could not be finalised until Brussels had approved the release of funds for the project.

The contracts would include an addendum comprising the timetable in document 4. The timetable would have a reference added to the effect that if there was a delay of more than one month in the signature (by CESD) of contracts then the dates in the timetable would move by an equivalent time.

8. Existing uses of business registers for surveys

Summaries of existing uses had been provided by nearly all participating countries prior to the meeting. These were brought together by the project manager in a presentation, attached as annex 1 of this report.

9. The seminar on the use of business registers

Document 8 outlined a possible seminar. A number of comments were made:

	The use of business registers and administrative data to supplement or replace statistical inquiries was an issue that was omitted.
	The UNECE meeting in Geneva had indicated that a seminar would be arranged to include issues surrounding use of administrative data and business register. The planned seminar on uses of business registers should try not to overlap with the UNECE seminar.
	The possibility of inviting participation of suppliers of administrative data (eg tax offices) was raised.
	Although the participation of business survey staff was important, there was seen to be little benefit in inviting national accounts staff.
	The first session on the context of the work on business registers should be very brief or omitted.
	A session on profiling of complex businesses would be useful.
	The seminar currently covered a lot of ground and it was important to get the balance between detail and range correct. It might help if some background material was provided in advance of the seminar.

Working group members were invited to make offers of specific papers that they might contribute by the end of May.

Although the location was not agreed, there were no objections to Luxembourg. Peter Struijs would provisionally book a meeting room in Luxembourg for February 1999. The date could not be set precisely because availability of rooms was always a problem.

10. Preparation of the next meeting

The agenda for the 2nd meeting would comprise:

	Update on technological issues (web site, PECO help and imputation software specifically)
	Planning for the seminar on uses of business registers
	Survey B3 progress
	Survey B4 preliminary planning
	Proposals for part 2 of the project
	Survey C publication and longitudinal data analysis (prior to the meeting data will be provided to participants)
	Initial discussion on population estimates

Countries were asked to consider if they could host the 2nd meeting. Offers were required by the end of May. Countries making offers should not enter into any firm commitments until advised by Eurostat.

Working group participants were asked to provide a description of their work area and where it fits into their national statistical office organisation, by the end of May.

Timing of the meeting was an issue. Poland was arranging a seminar on structural business statistics, probably in September. This was in the process of being re-scheduled from June. The Eurostat contact was Maria Pounder. Peter Struijs would check dates with her. It was also reasonable to avoid the dates of the roundtable meeting on business survey frames in Helsinki (end September). A date was set provisionally for up to 3 days in the week of 5 to 9 October 1998.

Following the meeting, Infostat requested that the next meeting should take place one week later (in the week of 12 - 16 October).

ANNEX 1

Existing uses of the business register

Summary of information from Working Group

members - 23 April 1998

Information sources

	PECO Panel - analysis of business register situation
	National accounts - register questionnaire
	SUPCOM - register co-ordination questionnaire
	One page summaries for this meeting

Main economic surveys

	Annual structural surveys
	Short-period production surveys
	Employment and wages surveys
	Producer Price Index collection
	Business tendency surveys

Surveys with special needs

	Agriculture
	R& D
	Tourism
	Foreign trade

	PECO Panel
	Register “proving”
	Economic census/ register survey

	Continuity - when existing register set up
	Coverage - private/ public, gaps of industries
	Quality of (administrative) sources

Business units

	Legal unit
	Enterprise
	KAU (Kind of Activity Unit)
	Local unit / Local KAU

Sampling/ analysis units

	Need for local units
	Local units for sampling
	Maintaining local units
	Sectoral/ regional analysis from enterprises

Survey issues (1)

	Sampling of smaller businesses
	Co-ordination of surveys
	Timing of selections/ despatches
	Tools available for sampling
	Tools for questionnaires

Survey issues (2)

	Identifying industrial sectors (NACE Rev 1)
	Importance of employment data on register
	Inclusion of small businesses in surveys
	Feedback from surveys
	Role of business registers unit

Statistics from the Register

	Monthly / annual structural publications
	Regional structural statistics
	Management of statistical surveys
	PECO Panel