The Business to Government (B2G) data sharing provisions included in Chapter V of the proposed Data Act are a measure that is necessary for the fulfillment of the stated goal of the regulation: ensuring fair allocation of value in the data economy. It is a crucial next step in the development of European policies that focus on the use of data for the public good. Until now, with Open Data policies, Europe has focused on enabling access to data generated by public sector bodies. With B2G provisions, similar rules will apply to data generated and held by private companies.
More fundamentally, B2G provisions, if made sufficiently robust, can shift the balance in the data-driven economy. Today, it’s the corporations that benefit from the value of data generated by all of us and collected by them. With these new rules, such data in particular – and commercial data more broadly – can be accessed and used for the public interest.
Yet the European Commission ultimately stopped halfway in fulfilling its own ambitions. Instead of proposing a structural, permanent approach for B2G data sharing and reuse, it proposes this mechanism as an ad hoc measure, to be used in emergencies and cases of special need. A more ambitious approach was on the table but was ultimately not chosen.
The Impact Assessment Report that the Commission published alongside the proposal is an important read that sheds light on the decision-making process. It shows how the impact assessment process is ultimately reduced to an economic cost/benefits analysis and market impact. Even though the accompanying study, prepared by a consortium led by Deloitte, signals that a purely economic comparison is not sufficient to choose between the two options that were considered.
It is a paradox that measures that flip the logic of the Digital Single Market, by introducing provisions that transfer data and associated value from the private to the public sector, are ultimately assessed in almost purely economic terms. Without new methods of measuring societal value and the non-economic impact of policies, European policymakers will continue to struggle with introducing society-centric policies.
The Impact Assessment: a look behind the scenes at the Data Act proposal
The document frames the choice that the lawmakers faced in terms of three Policy Options: a baseline option that is a “business as usual” scenario, and two new regulatory approaches of increasing intensity. The second option – one ultimately adopted in the Commission’s proposal – limits B2G sharing to an ad hoc basis, justified either by a public emergency or exceptional needs to use the data. This option also includes a business-friendly compensation model that includes a reasonable return on investment. The third, more ambitious option included a generic mechanism for requesting the reuse of business data by public bodies, for any “duly justified purpose”, without the need to demonstrate exceptional situations. Compensation was to be limited to marginal costs of providing the data, and furthermore, both public sector bodies and companies would need to designate a “data steward function” that would facilitate B2G requests.
Comparison between these three options is ultimately done based on a single metric – that of economic costs and benefits. So, the simplest answer to the question: why did the Commission choose Policy Option 2? is simple: because it will generate an estimated 273.1 billion Euro net benefits, as opposed to 221 billion Euro in the Policy Option 3 scenario. Even the impact on public administration itself – an important factor for policies that concern the public sector – is reduced to the impact on the administration’s efficiency, measured in economic terms.
The problem seems to lie not just with a lack of methodologies for measuring non-economic, societal impact. Also, the vision of B2G data sharing, with its huge potential to generate new services, innovations, and research, is lacking. This is illustrated by the fact that the document focuses on one simplistic case of B2G data reuse: that of increased availability of statistical information, which will generate spillover effects in the market. Europe’s ambitious vision of data for the public good should be based on a forward-facing vision of new data ecosystems. There are multiple policy tools available within the Digital Decade strategy to roll out a modern, even futuristic, data-driven public administration. But this will not happen if we face a failure of strategic imagination. The fact that a progressive policy package centers on 20th-century statistical institutions, with relatively limited associated data reuse, is a warning sign.
And there is an additional reason for this focus on statistics, which frames these institutions as the main beneficiary, and intermediary, of B2G data, flows. For this sector, separate impact assessment exercises have been conducted, providing quantitative estimates of economic impact. This example confirms the importance of establishing new measurement methods for different forms of value generated by data use. And obviously, novel ecosystems will not be measured as precisely as fields that have been established for over a century.
Policymakers struggle to quantify societal benefits
The precarious state of the evidence base is most visible on page 62 of the assessment, where the social and environmental impact of Policy Option 3 is meant to be analyzed:
“ […] a wider range of potential actors in the B2B context and a wider scope of applicability of the B2G provisions. This is expected to lead to substantial social benefits, although the support studies were unable to quantify these benefits”.
The assessment also fails to properly measure the value generated in the public sector. It states that while stronger B2G rules could benefit the public sector to a higher extent, this would be offset by a higher administrative burden and costs to low stakeholder acceptance. Again, burdens come to the forefront of the impact assessment, instead of a clear definition of positive impact. Tellingly, the benefit to the public sector is again reduced to “lower data acquisition costs” (related to a different compensation model). The assessment seems to lack the sense that the public sector can be generative, both in societal and economic terms.
This leads the authors of the assessment to state that “Overall, the evidence as to tangible improvement over Policy Option 2 in the B2G area is lacking”. But this does not mean that necessary evidence is not available. Rather, the evidence has not been sought. And the authors of the assessment admit, although this admission is hidden quite deeply in an impact assessment document, that:
“Our desk research showed that there is no well-established metric of the economic benefit of data sharing in general. This is also corroborated by interviews in this study and confirmed by meta-analysis: even participants in the data economy (i.e., those sharing data, and those receiving it) struggle to quantify the direct economic value of their data activities in terms of, e.g., turnover, profit, or efficiency gains. Even if such data were available, indirect value and externalities would not be appropriately considered (such as qualitative improvements in a product or service, new functionalities, better environmental performance, etc.). These are elements that no existing study has been able to quantify reliably”.
And a similar argument can be found in the accompanying “Study to support an Impact Assessment on enhancing the use of data in Europe,” carried out for the DG Connect by a consortium led by Deloitte. Its authors also believe that a cost-benefit analysis cannot be properly conducted to serve as evidence for the impact assessment. At the same time, they identify societal benefits that can result from adopting a stronger regulatory approach:
“While the quantification of benefits differ on a case-by-case basis and therefore a conclusion in terms of benefits is not possible to fully execute a cost-benefit analysis, as demonstrated during the evaluation of policy options 2 and 3, there are potential societal, environmental and economic benefits for private and public sectors (in terms of costs savings, efficiency gains) derived from a more structured and harmonised approach that incentivises business-to-government data sharing use cases”.
This is an honest appraisal of the state of evidence-gathering for data governance policies. One that should lead policymakers to carefully approach any simplified metrics of economic costs and gains. Yet that is exactly what has not happened in the impact assessment. Based on a purely economic cost/benefit analysis, stronger B2G provisions are seen as little more than additional burdens on market players. Especially since a market orthodoxy still underpins a lot of the thinking. For example, the assessment includes this statement, provided without evidence: “if companies are forced to share data in a wide range of situations under restrictive conditions, they are unlikely to make major investments in data generation, collection and handling”. This is a surprising argument, one that is repeated several times in the impact assessment – and one that we will, therefore, probably hear repeated in the policy debate. It argues that in our data-intensive economies, where in the last decades, almost any successful digital enterprise is founded on infrastructures for intensive data extraction, companies will stop collecting data out of fear of requests from public bodies to share them.
That there will be costs is obvious. Whether commercial data monetization will be impacted is less certain. But most importantly, we need to frame this as a trilemma, by putting into the equation also value generated through B2G data sharing. And we won’t be able to solve this problem without measuring the societal benefits of data treated as a public good and reused by the public sector. And if we listen to industry stakeholders, then we will simply need to give up on the idea of European common data spaces and data commons – on the basis that they would constitute a burden on full commercial exploitation of collected data.
Public interest: from a buzzword to a policy framework
In this analysis of the Data Act impact assessment, I have been focusing on demonstrating the lack of methodologies that would allow policymakers to measure societal impact related to new forms of data sharing. Having such methods would give us more balanced evidence, and, thus a better basis for informed decisions on the strength of policy interventions.
But the problem goes deeper, as there is also no conceptual frame that policymakers can use to define the public interest. We read in the assessment that:
“The concept of ‘public interest’ is generally recognized in EU legislation, but there is no harmonized definition. Member States have a wide margin of discretion in defining the exact meaning of ‘tasks carried out in the public interest’ or the related concept of ‘services of general economic interest’. Limiting the scope of B2G data sharing to exceptional situations makes the concept of ‘public interest’ as a requirement to determine what is covered less important. It is the exceptional character of the situation that will be the main criterion, rather than the notion of ‘public interest’.”
Seen in this light, the choice of an ad hoc solution allows policymakers to avoid the thorny issue of understanding and operationalizing public interest. But just like with societal impact measurement, this issue cannot be avoided, if Europe has the ambition of building a “data for public good” model.
The Commission is charting a new path with its high-level policy narratives. The Data Act, and the broader European strategy for data, are very different from the market-focused strategies of the Digital Single Market era. But the tools used to operationalize strategies seem ill-fitted to move away from the market orthodoxy of the Digital Single Market Frame and its singular focus on markets and economic growth. The same issue is visible in the European Digital Compass, where ultimately, the quantitative model reduces an ambitious vision of the twin transformation to a series of quantitative targets based on the basic vision that “more technology is good for the economy”.
Europe has a chance to build not just a data-driven economy (the prospects of which are not certain, in the face of competition from other regions), but a modern data-driven public sector, that uses “data for good.” But in order for this to happen, the European Union and European states need policy tools and frameworks that will make this vision operational. Without such tools, ideas like “public interest data use” will be limited to declarations of principles, but not translate into actual, observable uses of data.
