A lot has been written in the public press about fake news and the apparently magic abilities of novel, data-driven technologies to manipulate voters’ beliefs and, hence, the outcomes of democratic elections. The problem is rooted in recent developments regarding the information available about voters, means to provide information to voters, and the nature of information acquisition by voters.
These developments concern (a) data-driven voter research and the possibility of political microtargeting, and (b) news consumption of growing numbers of people using social media and news aggregators that obfuscate the origin of news, leading to voter unawareness about the news sender’s identity. Academic empirical research on the topic is scarce but growing. Theoretical research, however, that carefully analyses the ability of political interest groups to spread messages via social media to voters in a credible way, is lacking.
In a new working paper, “Big Data and Democracy” (joint with Freek van Gils and Wieland Müller), we provide a theoretical framework in which we can analyze the effects that microtargeting by political interest groups and unawareness of voters about the political position of a news sender have on election outcomes in comparison to “conventional” news reporting. The model framework does not only offer new theoretical insights, it also allows the theory-based discussion of policy proposals, such as to ban microtargeting or to require news platforms to signal the political orientation of a news item’s originator.
P.S. This is the sister paper to “Competing with Big Data” (joint with Christoph Schottmüller), which has had and is having quite some policy impact. Where “Competing with Big Data” studies effects of datafication on competition on data-driven markets, “Big Data and Democracy” studies effects of datafication on democratic elections and the process of political opinion formation. In both papers, we identify the core problem and analyze solution proposals, trying to inform policy makers.
Recently, I described the latest developments regarding efforts to establish a law mandating data sharing of information on users’ preferences and characteristics on data-driven markets (the economic rationale is here). The post ended with the statement: “We are actually in contact with that political party [in Germany, which made data sharing part of their platforms for the EU elections]. It will be interesting to see whether and how research can have real impact, which may impact the lives of virtually all Internet users (and beyond).”
Now we are a step further. Friedrich-Ebert-Stiftung, a foundation close to Germany’s Socialdemocratic Party, asked me “to develop some recommendations on how the idea can be implemented in legislative practice.” The outcome is a slim 15-page essay, titled “Competition Policy and Data Sharing on Data-driven Markets: Steps Towards Legal Implementation.” It tries to give (preliminary) answers to key questions:
- How to identify a data-driven market empirically?
- What information should be shared on which market?
- How can user information be anonymized and how can re-identification of individuals (technically or legally) be avoided?
- Who should share data?
- Who should have the right to get access to the shared data? At what price?
- How should data sharing be organized? What is the optimal governance structure?
Importantly, the answers given here are (informed) ideas, not fundamental research. But that is coming, too. The German Finance Ministry has commissioned a group of TILEC-researchers (combining expertise in economics, law, econometrics, data science, and consumer research) to study the key questions in the list above more seriously. Specifically, our tasks are:
- To develop an empirical test that could be applied by, say, a competition authority, which then could show that a market is “data driven”, or not. This would inform policy makers/regulators whether intervention by mandatory data sharing is innovation improving and, hence, positive for users on that market. We are also asked to apply that test to one industry, as a proof of concept.
- We are to develop a governance structure of data sharing, which implies to answer questions 2-6 listed above (and to show that it coincides with EU and German law).
The results of these studies are due in fall 2020. Hence, the saga will be continued.
The datafication of our lives, driven both by the availability of so much data and technological progress in artificial intelligence, is ongoing. One key element of this unstoppable process, which brings us many fabulous technologies but also exposes us to new threats, is the development of so-called data-driven markets, which has growing implications for all parts of the economy. In earlier work (especially here), we analyzed competition on data-driven markets and showed a very strong tendency towards monopolization/market tipping. Market tipping leads to lower innovation incentives of both the challengers and the dominant firm on these markets. Based on the analysis, we made a policy proposal, a specific kind of mandatory data sharing between competitors, and showed that this can avoid market tipping—at least in theory. More details are here.
Since 2016, policy makers have expressed grown interest in this research and in the policy proposal. Now, several European governments and the EU Commission have moved forward. For instance, the Dutch Economics Ministry recently wrote an open letter (in Dutch, English, French, and German!) to the Dutch parliament (guess why the many languages!), suggesting to significantly reform EU competition law such that competition authorities can already intervene in certain (data-driven, tipping) markets before it is too late, even if the dominant form has not abused its market power or cartelized the market, which current EU competition law requires. While this is already in line with our proposals, which have been expressed several times at that ministry, the letter picks up our data sharing proposal explicitly (on p.5 of the English version):
"Gatekeeper platforms may for example have access to certain data, such as specific consumer preferences, which other businesses need in order to compete, thereby making use of the platform unavoidable. In such cases the authority could force the platform to share the data in question with the businesses that require it under reasonable conditions."
Moreover, Sébastien Soriano, the head of France’s Electronic Communications and Postal Regulatory Authority (ARCEP), has put up five interesting proposals how (and why) to regulate big tech firms. His main reason (just as here) is to defend incentives of innovators, who cannot make business at all or face a very uneven playing-field, against big tech firms. Notably, one of his examples picks up our mandatory data-sharing proposal (the link to this website is in Soriano’s original):
"Data sharing mechanisms could also be used for competitive purpose. Economists have shown that, in data-driven markets, data sharing would allow us to regain our capacity to innovate: “Data sharing (voluntary, or not) eliminates the mechanism causing data-driven markets to tip.” This is no small issue: it would amount to protecting the innovation economy and ensuring that innovation is not the sole dominion of a small handful of people. This is why the idea is catching on more and more. One of Germany’s main political party’s actually made it part of their platform for the EU elections."
We are actually in contact with that political party. It will be interesting to see whether and how research can have real impact, which may impact the lives of virtually all Internet users (and beyond).
Update: In August 2019, the Dutch Competition Authority (ACM) published a short position paper supporting the above-mentioned open letter of the Dutch Economics Ministry. Specifically, they support “the introduction of an ex-ante intervention mechanism to prevent anti-competitive behaviour by dominant companies acting as gatekeeper to the relevant online ecosystem. We think introducing the mechanism by way of adding an extra tool to Regulation 1/2003, to be applied by the European Commission and the member states’ competition authorities, could be explored.” This is a great success as the ACM was rather skeptical about such ex ante measures, holding that existing EU competition law would be enough to confine dominant firms in data-driven markets, about a year ago.
A new initiative, Learn IOE, has been started at the University Paris Dauphine (where IOE refers to Institutional and Organizational Economics). The idea is to provide a stack of short video lectures about fundamental or current topics in IOE. In one of my contributions, I explain the concept of economic governance. In another one, I introduce a classification of economic governance institutions.
In short, economic governance studies how fundamental economic problems, such as contract enforcement, collective action, and the protection of property rights, can be tackled by careful institutional design. The categorization introduced in the video helps in a comparative fashion to identify the type of governance institution that is best suited to solve a specific application. In the next step, the (institutional and organizational) details of the optimal type have to be designed.
For an application of this concept to the cloud computing industry, see this! A previous paper that only did step 1 and characterized the optimal type of the governance institution for cloud computing is here.
In February 2019, the European Parliament, the Council of the European Union and the European Commission agreed on new European rules to improve fairness of online platforms’ trading practices. This nice fact sheet explains the rules in a nutshell.
Most of my recent research studies the consequences of datafication (the recent rise of big data and artificial intelligence) on markets, politics, organizations, and societies. This time it’s different. In “Data Science for Entrepreneurship Research: Studying Demand Dynamics for Entrepreneurial Skills in the Netherlands,” co-authored with Patricia Prüfer and just accepted for publication in Small Business Economics, we describe the most prominent data science methods suitable for research in the social sciences (here: applied to the domain of entrepreneurship) and provide links to literature and Internet resources for self-starters. We survey how data science methods have been applied in the entrepreneurship research literature.
As a showcase of how to use data science techniques, based on a data set of 95% of all job vacancies in the Netherlands over a 6-year period with 7.7 million data points, we provide an original analysis of the demand dynamics for entrepreneurial skills. We find that both entrepreneurial and digital skills are in increased demand for managerial positions in the Netherlands over the entire period 2012-2017. We also find (less surprisingly) that due to the hugely growing importance of datafication, amongst digital skills, those on ‘Digital transformation’ and ‘Big data and analytics’ are most valued by managers’ employers (less so for other professions). What is surprising, however, is that one could expect that demand for digital skills would increase most. Our empirical results, however, show the opposite: entrepreneurial skills were significantly more relevant over the six-year period studied. Moreover, the absolute importance of this skill type in managerial job vacancies has increased even more than digital skills.
Datafication has massively influenced processes within organizations, on markets, and more generally throughout society. Machine learning pushes the loop between data accumulation and innovation even further. The Tilburg Law and Economics Center (TILEC) and the Governance and Regulation Chair (GovReg) at University Paris-Dauphine | PSL Research University are pleased to announce the 5th Economic Governance workshop, which will take place at Tilburg University, the Netherlands, on June 6-7, 2019.
We now strive to stimulate the debate about the economic, political, legal, and social effects of big data and artificial intelligence. As a case of special focus, algorithm-driven platforms such as social media, search engines, and news aggregators have become dominant players in news dissemination. This has transformed the media sector and the way we think about democratic political elections and the legitimacy of those elections’ outcomes, with yet unknown consequences for our political systems and for many markets that are tipping towards the technological leader.
These developments challenge our rules of the game: are Western institutions, formal and informal, set up appropriately to ensure fair competition among firms, innovators, politicians, or political parties? What does it mean for competition law, privacy and data access laws, international treaties, election commissions’ procedures, and the codes of conduct on online platforms if most of us can be traced and monitored most of the time – but these masses of data can only be accessed, worked with, and potentially be manipulated by a few parties? Are we heading towards a future with virtually unbounded opportunities and progress for humanity – or towards a setting, where the state or large private actors control every aspect of life and the net profits of global technological progress are enjoyed by very few very rich and influential individuals?
The deadline for submissions is January 20, 2019. Details and further information are here.