Chairing the group at present is Boris Lorenc, a Senior Methodologist in Statistics Sweden, the Swedish national statistical institute. Boris talked to us to explain more on why ENBES was established and what they will be looking to achieve.
How did the ENBES project come about and what prompted the members of the group to recognise the need for this kind of collaboration?
It was a realisation that we in Europe lack exchange and networking in the field of methodology for business statistics, akin to that which exists for social statistics. In particular, we (Paul A. Smith (ONS) and Anders Holmberg (Statistics Sweden) and I) were prompted by a failure to generate enough European support for a proposal for hosting ICES-IV, the Fourth International Conference on Establishment Surveys (subsequently held in Montreal in 2012). However, in 2012 a European bid for ICES-V was successful, the conference is to be held in Geneva in 2016 with the Swiss Federal Statistical Office as its host.
However, beyond the conference itself, there was a clear need for more exchange on research ideas, practices, approaches and tools within the field of establishment statistics than what existed in 2008. Establishment statistics – about, or for, businesses, institutions, and other organisations – is in fact so much different from ‘ordinary’ statistics that is taught in survey methodology courses: in business statistics, we have a huge variation in units’ sizes, a very complex structure of units, increasing proportion of businesses being multinational, relatively large fluctuation in the target population (new businesses are emerging, existing are ceasing or integrated with other businesses, and so on), a response process and access to data that is much different than when surveying individuals, much better availability of advance information on businesses, and so on. These all, taken together, make business statistics rather distinguishable from the methodology of social statistics.
There is also the issue of response burden to businesses (it would not be unusual for a large company to need to provide corresponding to a full-time employee to the tasks of responding to official statistics surveys). Thus, how to feed some of this information back to businesses and make official statistics for them is yet another challenge.
From your initial meetings, what specific areas have you covered and what issues have you begun to collaborate on?
Our activity, which is completely based on volunteer work, is mostly reflected in biennial workshops and an online repository/wiki aimed at supporting collaboration between the workshops. We have held workshops, which stretch over 2½ days, in Stockholm in 2009, in Neuchâtel in 2011 and in Nuremberg in 2013. We have three main target groups: Academics working with methods for business statistics or whose fields of study may be relevant for business statistics, for instance organisational studies or econometrics, to give just two examples. Second, professionals working in official statistics – foremost national statistical institutes, but also other producers or users of official statistics, like central banks, and similar. Finally, organisations from the private sector that collect data from businesses, produce, or use business statistics, including but not limited to business associations and media.
We have from the outset stressed the need to cross existing boundaries (in statistics production, academic fields, countries, between private and government sectors), seeing benefits in learning and being inspired from what others in the adjoining field are doing. We used the phrase ‘cross-fertilisation’ a lot in the beginning. As an example from the statistics production, there is a step after data collection called editing. One goes through the collected data, identifies potential errors and in some cases goes back to the reporting business and verifies correctness of the data item.
The editing process may require considerable production resources (30% of the whole data collection costs would not be unusual), plus generates additional response burden at the businesses. Thus, understanding the causes of errors and correcting them (in addition to perhaps correcting the consequences of errors in the form of erroneous data) is important. Often, the form used for data collection needs improvement: it is not obvious in which units are the data to be provided, what exactly is meant by some of the terms used in the form, and so on. But, usually there is no feedback from data editing people to others who are designing questionnaires.
Do you think greater co-operation in projects such as this will help European economies to further integrate into a more cohesive unit?
Independently of our initiative, in 2008 a large Eurostat project on improving economic and trade statistics – called MEETS – was initiated, which is still on-going in the form of a number of more specific projects. Further, between 2010 and 2013 there was an FP7-financed project on business statistics called BLUE-ETS. These activities definitely led to spreading of high-quality methods and practices for business statistics. While financing such projects is likely correlated with an integration of European economies, I see that question as a separate – and political – matter. As appropriate as possible theories and methods for business statistics should be used independently of how integrated European economies are. For business statistics, we also have a lot to learn from countries outside Europe, like New Zealand, Australia or Canada.
What do you think can be achieved by the project?
ENBES’s goal is to continuously support and encourage work on improving theories, methods and applications for business statistics production over professions, academic fields and countries, wherever this may lead us. I believe that there is a need for such open exchange and learning.
Image courtesy of Monica Holmberg.