According to Spanish regulation, every public administration at the national level needs a part on its web page where it publishes open data that is also federated to datos.gob.es. Furthermore, it forces all organisations to develop an open data strategy and establish a designated unit within each ministry in charge of open data, thereby creating a spread out network of “advocates” for open data. However, these ‘open data officers‘ do not have any significant decision making power over which data to publish. These decisions remain with the data owners, specific departments within the ministry that generate the data in their regular work.
What can be observed is a conflict within the public administration itself. On the one hand, there are employees whose job is to actively push the topic of open data. On the other hand there are those employees of operational departments who ultimately have factual decisive power over the data that is supposed to be provided. Often their intention is to shield valuable, sensitive data. Heimstädt describes this problem as one of unequally distributed benefits that relates to the characteristics of the data (Heimstädt et al., 2014). Controlling the possible usage of the data is on the one hand undertaken to shield politically sensitive data. Also, shielding data has something to do with covering up the poor quality of the data itself.
The situation then arises that “you can have an open data catalogue with thousands of data sets, it looks pretty good, lots of data are available, but if you start to work with one single data set you see that you just have rubbish.” (open data advocate and consultant, Spain). Without satisfying neither the initial political idea of open data, nor potential users, this practice nevertheless provides open data officers and strategists some kind of legitimacy, as well as retains control of the data to those public servants who always used to have it.
In Germany, the situation appears to be quite similar. Over all three subsequent open data projects in Munich the decision whether to publish data, which parts to publish and how to publish it is left to the administrative departments that generate the data in the cause of fulfilling their public task. These departments have very few incentives to publish data as open data. The strategy units in charge of open data and the political level share the praise for the launch of a platform, while blame and shame are largely left to data contributors, when false data are published. Praise for contributing significant data presupposes political attention. There never was a strong demand from the political level to publish open data, barely holding out the prospect for praise when making their data available. Rather, they take a number of risks when publishing open data.
The risks involve the above mentioned liabilities, if the data includes copyrighted third-party content as well as possible flaws in the data. Thus numerous legal questions arise, giving special prominence to legal departments. Regarding the risks of possible privacy infringements, data protection officers are regularly involved in the deliberations. The risks also pose a potential embarrassment for public administrations that generally have a low tolerance of mistakes. Furthermore, fear is prominent that transparency might force politicians’ and administration’s hands to tackle societal problems that are either hard to solve or require massive financial expenses. These consequences might arise from a clearer picture of a current situation as well as different appreciations of the problem and interpretation of the data.
Also open data users have so far not proven or even aroused the perception among data owners that they are capable to deliver meaningful results. The number of actual users who visibly make sophisticated use of open data is perceived as low. When users file requests for specific data sets, for practical reasons they can often not make specific statements what the data will be used for and what the impact will be, for they do not know the content and quality of the data in advance. What remains is the promise of sophisticated and powerful ICT capable of processing, mashing up and analysing large amounts of different kinds of data.
Somehow exceptionally are those parts of public administration that are more used to publishing their data, though not necessarily to the general public or for free. Here ICT systems are ready-made to prepare and share data, e.g. geographic and meteorological data. Also, there is a history of standardisation across administrations and jurisdictions that lead to nation-wide or even European standards, e.g. the INSPIRE directive for geographical data.
However, these administrations often rely on their data as a source of revenue that would need to be compensated. Furthermore, there is staff in those departments that prepares data and handles the distribution and financial transaction processes. These positions might become redundant, when data is completely made available as open data. Thus, efficiency gains are not accepted as an argument. Therefore, benefits – if feasible at all – are unequally distributed.
Nevertheless, open data is perceived as a noble endeavour that lends the appearance of modernity. Thus, politicians pay lip-service to open data without establishing a framework in which it can truly thrive. As a consequence, window-dressing strategies currently best align actor’s interests: Administrative departments provide data that is available in a structured format, at a fairly good quality level, not obviously sensitive, requires little maintenance and thus little effort to be published. This can be termed the “availability approach” to publishing data (Hunnius & Krieger, 2014).
Leaving the decision about a specific dataset to the various ministries creates a spread-out network of data generators. These often have some topical understanding of what the data is about, but not necessarily a profound technical understanding of how open data are used, what requirements for use are or what facilitates their use. Also, the individual person might become some sort of internal advocate for the topic he or she is responsible for, but they mostly lack decision-making powers which is internally further dispersed to heads of departments. Thus, under such circumstances the internal open data advocates largely depend on the leadership from the organisation’s senior level to encourage and direct departments to make data available, since policies are often discretionary and not mandatory and leave considerable leverage to the agencies (see also Shkabatur, 2012).
“In general, it is the availability of potential data. This is still a point that data are selected based on how easy it can be made available and less based on its usefulness.” (interview participant, local level Germany)
Stakeholders and their exemplary interests in decisions about publishing data sets
|Data owners – heads of departments||Furthering their professional cause; appearance of modernity, retain independence and influence|
|Open data officer||Persuade data owners to make open data available and advise on decisions|
|ICT/open data strategy unit||Maximise the number of data sets on the jurisdiction’s portal|
|Legal department||Develop policies and guidelines that conform with the legal framework and stand the tests in court|
|Data protection officer||Ensure data privacy while also ensuring administration conforms with FOI legislation|
|Open data users||Showcase the benefits of open data in order to push governments to publish more open data|
References can be found here: OpenDataMonitor Project – Shared References