Page 308 - Special Topic Session (STS) - Volume 2

STS496 Roeland B.
course, in a context of competing demands within government official
statistics needs to get their voice heard. It is therefore important that we -
statistics producers - should not refrain from pointing out the value of our
work. This means that we have to be able to prove the relevance of our
statistics to policymakers within government who ultimately decide on the
resources we work with. We have to be clear about the value and usefulness
of official statistics for government and for society at large. We should present
positive arguments for receiving sufficient resources in order to be able to
develop and maintain a strong official statistics infrastructure.
However, the often hard-won resources available within an official statistics
system should also be deployed in the most efficient way possible. In
programming the work of an official statistics system, we should set clear
priorities for what is needed, but we should also stop producing statistics that
are not or hardly needed, in order to release funds for the production of new
statistics with higher priority.

2.3 Mandate for data collection
The data sources that can be used for the production of official statistics
have been changing in the past decades. Traditionally, official statistics have
always strongly relied on censuses or surveys among persons, companies or
institutions to generate data. This classic model of censuses and sample
surveys is increasingly replaced by strategies in which two other data types
play a increasingly important role in addition to censuses and surveys:
administrative data held within government and - more recently - big data.
This evolution is sometimes described as an evolution from ‘designed data’, in
which the statistical expert can autonomously determine the data collection,
towards ‘organic data’, in which statistical expertise contributes very little, and
already existing data are relied upon (Groves, 2011).
This evolution has had significant consequences for the data availability for
official statistics. With this shift the locus of data ownership has moved away
from statistical offices to a range of diverse players outside the statistical
office: other government departments or agencies, and increasingly private
sector actors. In some countries recent legislation recognises this shift, and
provides for compulsory access to these data by the statistical office (e.g. UK
Statistics Authority, 2017).

3. The organisation of statistical systems in democratic societies
As explained before, official statistics systems are by default part of
government. However, there is some debate – and difference in practice – as
to which branch of government an official statistics system should belong to.
An in-depth review of governance models for official statistics is not within
the scope of this paper, but it appears that in the majority of countries the

297 | I S I W S C 2 0 1 9