Back to Top

How is the CSO doing?

Your feedback can help us improve and enhance our services to the public. Tell us what matters to you in our online Customer Satisfaction Survey.

 Skip navigation

Enterprise Statistics Liaison Group (ESLG)

Held at the Central Statistics Office (CSO) Ardee Rd., Dublin on 4th March 2016.  

 

Agenda:

1. Introduction by chair

2. CSO updates on:

     2.1 Access to research microdata files

     2.2 EU matters, including SIMSTAT

     2.3 Structural business & business demography data revisions - data coherence project

     2.4 National Employment Survey / SESADP, Structure of Earnings Survey, Administrative Data Project

     2.5 Access to Finance survey and release

     2.6 Addition of balance sheet questions to survey forms 

3. User Forum  

 

1. Introduction by Chair

Chair welcomed us all to the meeting and asked everybody in the room to introduce themselves.  There was a brief summary of the last meeting and outline of the agenda items of this meeting.  

 

2.1 Access to research microdata files (RMFs)

An update on access to Research Microdata files since the last meeting. 

Background

Currently, a project team at CSO is working on delivering remote access to researchers. The 5 SAFES as listed have to be considered as these underpin CSO’s present procedures for research access:

  • Safe data (documented and minimised for privacy purposes)
  • Safe researchers (accredited researchers)
  • Safe projects (public interest, ethical etc.)
  • Safe outputs (non-disclosure aggregates)
  • Safe setting (data processing is done in a secure environment). 

The plan

Building a new researcher network at CSO that can provide some new benefits (i.e. shared folders in a project team, backups, and virtual desktops providing remote access). This researcher network will provide a modernised safe setting for research projects. 

Currently, the proposed software list for the standard virtual desktop is:

  • R
  • STATA 14
  • Libre Office
  • Matlab.

Feedback on the proposed software suite was invited. The software package will be on the server in the CSO so there should be no cost to the user.

The plan with limited resources is to set up a DATA office in 2016 which will take over administration of research access in 2016.  The aim is to reduce the “red tape” and to have a stronger co-ordination and management of research projects. Each research project will be allocated a single person to act as a liaison with the project manager.

Current state of play in Remote Access project

At present the CSO is working on acquiring the necessary infrastructure to support remote access.  If all goes to plan the rollout of remote access would be in 2017.

Administration of RMFs and one year terms

Decisions on applications for research access are taken on a case-by-case basis bearing in mind Section 20(c) of the Statistics Act

any other person authorised in writing by the Director General to perform for a specified period particular statistical analysis which may necessitate access to data collected under this Act”.

To comply appointments as an Officer of Statistics for research purposes are made for no longer than one year and for one specific project.

Managing expectations

It is anticipated that the demand for researcher access and the complexity of that demand will continue to increase. Due to resource constrains some valid projects may not be realised.  

Special issues to be considered in relation to administrative data

RMFs from administrative data sources (about persons or enterprises) are considered a higher privacy risk. Example the NES has nearly all employees in the State so risk of disclosure is greatly increased.  Additional measures may include on-site access only and access to sampled rather than exhaustive data.  These measures are not finalised.

There was some useful discussion after the presentation in particular:

The length of time for a decision on whether access will be given was discussed.  The CSO is very mindful that the wait time is frustrating to users and is setting up the Data office and procedures around it to cut out the internal “Red tape”.   

The time period of one year for access was raised by members of the group.  One year is chosen as it is felt any longer would not be considered specified, as per the Act it is also felt for most research projects that 12 months is sufficient.

The access as per the Act can only be for a “particular statistical analysis” so broad access cannot be given.

Normally once data is published anyone can ask for access to it via RMF.  Generally (outside of the current coherence project) Structural Business Statistics (annual statistics) RMFs are available 18 months after collection. 

2.2 EU matters, including SIMSTAT

A new consolidated Regulation on business statistics, called the “Framework Regulation integrating European Business Statistics" (FRIBS) is being prepared by Eurostat.  This will consolidate and introduce new data collections in the following areas:

  • Short-term statistics (monthly and quarterly indicators)
  • Structural business statistics
  • Business demography and business register statistics
  • Inward and outward foreign affiliate statistics (FATS)
  • R&D and Innovation
  • Information Society Statistics
  • ProdCom
  • International trade in goods statistics (ITGS) -  SIMSTAT

SIMSTAT

This is Eurostat’s preferred approach to simplifying the present Instrastat system and reducing burden on traders. This is the most contentious item in the proposed FRIBS Regulation.

The proposal will involve

  • The mandatory exchange of trade micro-data between Member States
  • Exporters will give Importers details at product level therefore in theory reducing the need to collect import data

CSO concerns:

  • transmission of commercially-sensitive Irish micro-data to other organisations in the EU including those with dual mandates,  e.g. customs and statistical,
  • Technical issues -  data being processed by multiple organisations instead of one

Quality and timeliness dependant on data flows from 27 MS instead of from one internal source

  • Not confident that there will be a burden reduction for importers only see an increase for exporters as countries like Ireland may still need to collect import data for data quality purposes
  • Impact on Irish business statistics if the introduction of this SIMSTAT proposal impacted on response levels from key enterprises in particular. Irish top 50 exporters account for 2/3 of Intrastat goods exports and 40% of Imports plus very large impact on National Accounts, BOP, Business Statistics

CSO therefore will continue to support the collection of import data through the Intrastat survey for the foreseeable future, regardless of whether SIMSTAT is enacted or not. We believe that this is necessary in order to maintain quality and timeliness of statistics.

Eurostat currently aims to present final proposals on FRIBS and SIMSTAT to the European Statistical System Committee (ESSC) in May 2016, after which it may become a formal Commission proposal for new legislation and then the decision making process in Council and Parliament will begin.

The CSO will also consult a number of individual enterprises about the SIMSTAT proposal.

The topics brought up during the discussion following the presentation were

Burden

The CSO does not foresee a reduction in respondent burden as do not think it will be possible to drop the collection of Import data from Importers as we are so dominated by a handful of companies.  There is a concern over quality and timeliness due to the number of sources that now would be involved and the data checking that would be required.  In fact an increase is foreseen as every exporter would have to give the importer details and the detail required would be at transaction level CN code, destination tonnage, partner ID, VAT number. 

Risks

Sharing of data across 28 states and the Commission increases the risk of disclosure.  The dual mandates of some organisations in other MS would also be an issue.

Political / Legal

A qualified majority is required to pass SIMSTAT into law, and will be dealt with through the usual channels..

Eurostat

There was a pilot of SIMSTAT in twenty countries and based on this Eurostat thinks it works.  The exchange of data worked but there are quality issues.   In the pilot France and Italy said there was a good match, although these countries already have some data sharing /matching on a bilateral level.  Eurostat believe there will be an overall respondent burden reduction. 

General

There have been as yet no discussions with individual companies. This is the first discussion with collective bodies. This only affects intra EU trade. There will be no change to the collection of data concerning extra EU trade.

Currently the Trade statistics are published within two weeks – relying on information from 27 countries could impact on this timeliness.

2.3 Structural business & business demography data revisions - data coherence project

While CSO cannot speak about the specific data changes as it is not finalised yet, the broad outline is a follows:

The data changes:

  • Business Demography/Register – finalise the active population of enterprises
  • Clearer definition of “active enterprise” now – remove dormant enterprises & remove process thresholds concerning what enterprises to count as active
  • Reconcile new SBS GVA estimates to National Accounts GVA estimates 

Other data changes

  • Non response treatment – use Revenue data more
  • Large Cases Unit review of some enterprise data
  • Aircraft leasing new series – include in SBS Services
  • R&D – capitalisation of R&D (adding to GVA)

Note: 

  • Net effect will be increase in active enterprises BUT low impact on GVA
  • Main impact on SBS data, minimal on National Accounts data


New Publication Tables, NACE & Size Classifications

New size classification across the three SBS surveys based on number of persons engaged;

  • Micro: 0 to 9 persons
  • Small: 10 to 49 persons
  • Medium: 50 to 249 persons
  • Large: 250 or more persons

CIP changes

  • Size classification as above
  • reduce the number of tables published from 18 to 8 – use Statbank instead
  • NACEs will not be grouped together to overcome confidentiality issues – there are problems associated with ad-hoc groupings

ASI and BCI

  • Change in size classification as above
  • No change in number of tables published
  • No change in NACE classification

Where are we now?

  • Creating new business register frames approximately one week from finalisation
  • Create new SBS & Business Demography time series for 2008-13 and test and analyse the data  there will be a break in time series                                   

 

                       Eurostat Transmission & Validation Clearance
                       National Publication & Statbank Tables

 

  • RMF datasets - New microdata files available for CIP, ASI and BCI for years 2008 – 2012 & files for reference years 2013 and 2014 also to be made available for the first time.
  • Doing the above work at same time as ongoing 2014 work but intend to finalise as soon as possible

The main items that were discussed after and during the presentation were:

Active Enterprise definition

The previous/current definition for an active enterprise, while using the presence of turnover or persons engaged as an indicator of activity, also allowed for any national method NSIs  thought relevant to use as an indicator of activity.  This subjective element in removed in the proposed FRIBS.  An enterprise will be defined as being active if it has

-           turnover    AND / OR

-           persons engaged AND /OR

-          Made  investments

Effect on bottom line

The finalisation of a new active enterprise population is likely to have a relatively small impact on overall GVA.  The other data changes -  Large Cases Unit review of some enterprise data.  Aircraft leasing new series and the capitalisation of R&D will have a greater impact on GVA. 

SBS v National Accounts

National Accounts and SBS measure different concepts and so have bona fide reasons for the differences in GVA arising from the two areas.   The publication of the new SBS series will include an explanation of the differences between National Accounts and SBS.

The changes to the publication / release

The number of CIP tables (18) is a legacy of the paper publication.  There will be a reduction in the number of tables based on the number of hits each had on the CSO website.  The less popular tables can be run off as a request/query. 

NACE categories will not be grouped together to overcome confidentiality.  If a NACE is confidential it will be not combined with other NACE groups.  Often the NACE groupings had no logical association and could change from year to year.

RMFs

There was a discussion concerning the RMF implications of the new SBS series. The users present at this meeting detailed that they would have no issue with the new datasets based on the updated series as long as the ‘old’ datasets were also available as it is very important that research can be duplicated.

Further progress on this project, including indicative notice of publication, will be communicated to the group.

Back to Top

2.4 National Employment Survey / SESADP, Structure of Earnings Survey Administrative Data Project

This item concerned how to meet data needs from administrative data and modelling rather than using the more traditional survey based approach. This has the potential to generate significant cost savings and timeliness improvements. The presentation detailed how the following datasets could be linked and modelled together in order to meet user needs in this general earnings area:

  • Administrative: P35L, DSP
  • Statistical: Business Register, EHECS, Census 2011, QNHS, SILC 

The finalised data would meet the EU regulatory requirement and would also provide RMF datasets for research purposes.  The intention is to submit the results to Eurostat by mid 2016 with a national publication after.  

The main points discussed following the presentation were:

The addition of the CSO  data to the Revenue & DSP data can add other variables including NACE, Gender, Education level etc.

As per the earlier discussion on RMF access generally, only a sample of this finalised dataset would be available for RMF purposes.

Peer Review

The CSO commissioned a peer review from the UK Office of National Statistics (ONS) to validate the methodology being used and identify any gaps or issues which need to be addressed.  The ONS Deputy Director of Labour Market Statistics visited the CSO for two days in February and the resulting report provided numerous helpful suggestions, in relation to specific variables and in relation to the processing and publication plan for the Structure of Earnings Statistics.

2.5 Access to Finance survey and release

This had been published on the CSO website the day before.  http://www.cso.ie/en/releasesandpublications/er/atf/accesstofinance2014/

Survey details:

  • SMEs only – for the first time all SMEs surveyed
  • Sample size of 7,000 (x 2 of previous sample size)
  • Response rate of 27%  (on a like for like basis we have a higher response rate than ATF 2010)
  • Series break – can’t really compare ATF 2010 & 2014 due to the inclusion of new questions in ATF 2014 and due to the inclusion of MICRO enterprises (employing < 10) in ATF 2014
  • Weighting effect – the MICROs can tend to dominate the total for SMEs given that MICROs a/c for > 90% of SMEs

The publication was very much welcomed and the CSO was thanked for the work involved. 

The other points mentioned were:

There were too many changes to the methodology and questionnaire to compare with the 2010 survey.

Users showed great interest in the publication and would like it to be a regular survey.  It was pointed out that making major changes to the survey would make comparisons over time difficult.

2.6 Addition of balance sheet questions to survey forms

Background information – certain balance sheet questions were added to the SBS e-forms for reference years 2014 & 2015.

Response Rates for 2014 for the balance sheet questions:

CIP 2014 Almost 16% made an online response (504 in number but 74 of these left all the balance sheet questions blank).  The 430 balance sheet responses are breakdown as:

–     Micro (0-9) was 5 responses                           Small (10-49) was 164 responses

–     Medium (50-249) was 197 responses             Large (250+) was 64 responses

–     74 responses to the CIP 2014 online form There was 504 online responses out of a total post out of 3,210 enterprises (15.7% response rate)

–     There was 2,078 responses in total out of a post out of 3,210 enterprises (64.7% response rate)

ASI 2014 had 1,918 online responses (almost 11% of total), but 313 left the balance sheet questions blank, with the following breakdown of respondents:

–     Micro (0-9) was 645 responses                                   Small (10-49) was 613 responses

–     Medium (50-249) was 280 responses             Large (250+) was 67 responses

BCI 2014 had 613 online responses (c. 13% of total) but 446 left balance sheet questions blank, with the following breakdown of respondents:

–     Micro (0-9) was 58 responses                         Small (10-49) was 71 responses

–     Medium (50-249) was 34 responses               Large (250+) was 4 responses

What Happens Next?

  • Access to the data will be made available pending approval of Officer of Statistics applications. Data for reference year 2014 will be available once data is published (approx. Q3 2016).
  • Data collection of balance sheet questions to be repeated for reference year 2015.
  • Decision on the inclusion of the questions again for reference year 2016 to be taken before the end of this year

The main discussion points were:

It was asked does the CSO envision the balance sheet questions being included on the paper version of the survey.  The CSO wants enterprises to move towards eforms.  There is a problem with the transition for some companies who keep a paper record of the previous form and this helps them fill out the current year form on paper.  The paper survey is already 8 pages long so it is not considered a good idea to make it any longer and so there is likely to be no room to add the balance sheet questions to the hardcopy form. 

3. User Forum 

Tara Davis (CSO) chaired this part of the meeting.  She introduced herself as this was her first time at the group as she has become Head of Short Term Statistics Division. 

She announced to the group that CVTS survey is coming up and asked the different enterprise bodies to let their members know.  It is a Eurostat mandatory survey which should be carried out every 5 years.  It is 8 pages long and will involve many areas of an enterprise but mostly HR.

Some User suggestions raised at the meeting:

  1. Productivity measurement

There are a number of international productivity measurements available, for example from the OECD.  Users queried whether CSO would be entering this space also, and perhaps provide a productivity release like the UK do?  CSO undertook to raise this issue with our National Accounts Division.

       2. Entrepreneurship indicators

An outline of available entrepreneurship indicators was provided.  It is likely that future indicators also arising from FRIBS.  At the next users meeting, CSO will outline the changes / requirements under FRIBS in this general area and this may overlap with user needs. 

       3. Payment performance

A query was raised whether CSO could conduct a survey to measure the above – essentially to determine the number of days enterprises wait for payment from other enterprises. 

       4. Sub regional needs

There is an appetite for more data here.

       5. More detailed user needs engagement

It was accepted that it would be useful to conduct a more in depth focus session on user needs arising in the short to medium term.  CSO will organise this meeting. As an input to this meeting, CSO will provide an outline of the enterprise statistics available now and the future enterprise data collections as envisaged under FRIBS. This can be a starting point for an exploration of Irish enterprise data needs.

Meeting conclusion

Keith McSweeney of CSO summed up the meeting and thanked everybody for their attendance and participation.  The CSO is determined to meet National user needs not just Eurostat requirements and look forward to a really positive engagement in the near future.  

 

Back to Top