Rainer Kattel and I have written a paper that will be contributed to an edited book sometime next year about policy success stories. We have decided to publish a pre-print version to get the word out early. Our paper titled “Estonia’s digital transformation: Mission mystique and the hiding hand”.
The original blog post about the paper appeared on the Medium blog of @IIPP_UCL:
In 2017, The New Yorker published an article titled ‘Estonia, the digital republic’ and subtitled, ‘Its government is virtual, borderless, blockchained and secure. Has this tiny post-Soviet nation found the way of the future?’. This summarises the buzz around Estonia’s digital government: from the outside, at least, it is seen as a major success and has been lauded in mainstream media such as the Financial Times, New York Times and Forbes. Given this success, it is somewhat surprising that this narrative, and its main drivers, has not been actually properly documented in academic research. In our new working paper, we do exactly that using interviews with all key architects of Estonia’s digital government.
What were the main drivers behind its digital transformation? And perhaps most importantly, what does Estonia’s digital transformation tell us about the future of (digital) governance?
What does success look like?
The Estonian e-government infrastructure and its success rest on two main pillars, both introduced in 2001, which essentially create digital access to state and digital citizens: the data infrastructure x-road and a compulsory national digital ID. X-road is an interoperability platform for existing decentralised databases and a data exchange layer that can be used by public and private sector actors. It is independent of platforms and architectures and provides secure interoperability for data exchanges and identification of trusted actors in digital service delivery. The digital ID makes it possible for citizens to be identified digitally and to use digital signatures. Together, x-road and the digital ID make it possible to digitally sign any contract, access essentially any public service, order prescriptions, file taxes, vote and so forth.
More than 2,300 public and private services use x-road, and the digital signature has been used almost 350 million times by Estonia’s population of 1.3 million. The digital ID penetration is close to 100%; 30% of votes are cast digitally (in both local and national elections); almost all personal income tax declarations and medical prescriptions are done online, and most medical records held by hospital and family doctors are accessible online. The Estonian government claims that its digital infrastructure has led to annual savings of about 2% of GDP and more than 800 years in working time for the public and private sectors.
According to the EU’s Digital Economy and Society Index (DESI), in 2017 Estonia was the leading nation in Europe in digital public services, although in 2018 it dropped to second place. However, in most other e-government rankings Estonia‘s digital success is less evident. In DESI’s overall ranking, Estonia is in ninth position for 2018 and, according to the UN’s 2018 e-government survey, it is ranked only 16th globally. This cognitive dissonance — high praise and leadership position in global news outlets versus relatively middling rankings in overall digital transformation indexes (for more discussion, see Drechsler’s contribution in this new book) — reflects the nature of Estonia’s digital success: Estonia is ranked high for its digital public service infrastructure, which is universally available and mandatory, and an integral backbone of public service delivery. Estonia’s digital success, however, is not about other digital offerings such as digital democracy, citizen engagement or digitally transforming public services such as the welfare state.
The specific nature of Estonia’s digital achievement and at the same time disconnect between technological infrastructure and degree of digital penetration is often overlooked in international coverage. As Estonia’s digital government came to be focused on the x-road, Estonia has effectively created its own legacy system — a move that the initial thinkers wanted to avoid. Indeed, in the early 1990s, the focus was as much on secure digital infrastructure as it was on advancing societal goals through digital means. Yet, in particular throughout the 2000s as Estonia blazed through an economic boom that created increasing inequalities in its wake, the evolution of digital government seems to stalled.
In some ways, Estonia’s digital government has been caught up in its own success: in 2014 Estonian introduced e-residency programme that opens some digital public services (establishing a company, paying taxes) globally. While more than 40,000 e-residents have signed up — most recently, the pope was gifted e-residency — and more than 4,000 companies have been created by e-residents, the programme has also faced domestic criticism as a something of a show-off that enables money laundering. (Similarly to India’s Aadhar, Estonia’s digital infrastructure faced constitutional court challengesthat were related to e-voting; as in India, Estonian court ‘sided’ with digital solutions.)
While many digital services have brought efficiency gains to citizens and businesses in Estonia, citizen satisfaction with crucial services such as healthcare and education has remained low. As an example, according to OECD rankings in 2014, Estonia ranked fifth from bottom in satisfaction with health services (in 2007 it listed the lowest) and second from bottom in education system satisfaction (in 2007 it ranked third lowest). Further, Estonia performs poorly on some critical social indicators. For example, in 2016 it had the highest gender pay gap in Europe and a higher than EU average Gini index. While citizen satisfaction is not the only measure of the quality of public services — and digital infrastructure is only one component in the provision of sophisticated services such as health and education — it is indicative that there is a little-measured improvement in the provision of core public services.
How did the success come about?
Perhaps most surprisingly, Estonia has never had a central office for digital transformation, such as the UK’s Government Digital Service (GDS), even though such a central agency was initially envisioned to manage (among other things) vital public registries. Estonia’s digital transformation has been an extended and on-going process over three decades, starting in the early 1990s, when Estonia regained its independence and continuing to the present day. Much of this process has been ad hoc and informal. For example, many strategic policy documents for digital transformation have followed the rhythms of European (structural) funding periods rather than responding to domestic challenges and planning processes. Similarly, various overlapping and mostly self-managed public-private networks have provided the informal dynamic capacity and capability for change, few of which have been institutionalised or formalised. And above all, the process relied initially on what Albert Hirschman has called the principle of the hiding hand: policymakers push visionary changes without anticipating all the challenges and risks involved upfront, an approach that sometimes results in unexpected learning, creativity and — in this case — success.
The naiveté and enthusiasm of the hiding hand that propelled the initial ‘crazy ideas’ of the early 1990s became ingrained in Estonia’s digital policymaking culture. As Mart Laar, prime minister in 1992–1994, and 1999–2002, and perhaps the key politician in this story, once said: “I was 32, I was young and crazy, so I didn’t know what was possible and what’s not, so I did impossible things.” (As he told us, for him the e-residency programme is simply not crazy enough.) Estonian digital government came to rely heavily on such charisma and hacker mentality.
The hiding hand was propelled by another rather simple phenomenon: envy. The success of Finland and Nokia became one of the guiding political reference points in the 1990s: ‘What is our Nokia?’ asked Lennart Meri, Estonia’s first president after Estonia regained independence in 1991, a question that has remained a popular catchphrase. Importantly, the confluence of Soviet industrialisation with ethnic tensions (mass immigration in post WWII era was related to rapid industrialisation during this period) explains why Estonia, despite being one of the most economically and technologically advanced countries within the Soviet Union, chose not to upgrade the inherited economy and instead sought to do something completely different. As an emerging general-purpose technology, ICT offered an almost perfect solution, particularly given the availability of R&D skills in this sector. ‘The digital’ thus came to express Estonia’s — or more precisely, its elite’s — ambitions and explains why, to this day, the ‘digital elite’, with some rare exceptions, is almost all ethnically Estonian.
There were three further vital contextual factors:
- While the Soviet legacy had left Estonia with an outdated industrial structure and widespread aversion to industrialisation, it also left Estonia with a wealth of R&D talent in ICT through various Academy of Sciences institutes, such as the Institute of Cybernetics (established in 1960, its spin-off Cybernetica AS developed x-road and e-ID) and other similarly highly advanced academic institutions. As their funding collapsed in the early 1990s, much talent poured into emerging private companies, in particular, various joint ventures with Scandinavian telecoms and other companies.
- Estonia has geographical proximity to Scandinavia, and in the 1990s the Nordic countries had one of the fastest developing telecommunications sectors globally. Opening up both policy-making processes (through advice and joint ventures) and markets (through privatisation and regulations) to Scandinavian partners brought know-how and investment.
- Estonia is a small country with a high population concentration in its capital city: almost one-third of its 1.3 million inhabitants live in Tallinn. This clustering of population facilitated agile networks that were able to gain quick and lasting political support, and which required low initial infrastructure investment.
Within this ideational and historical context, three critical features of Estonia’s digital transformation, still dominant today, emerged:
- future-oriented and almost utopian solutions — the realm of ‘crazy ideas’;
- public digital architecture that is universal in nature; and
- decentralised digital agendas (including databases) of line ministries and agencies.
Silicon Valley of digital government?
Estonia’s focus on ICT as general-purpose technology has proved to be one of the critical success factors as it enabled Estonia to create a digital infrastructure that is universal in nature. Yet by relying on decentralised and mostly informal networks to build this infrastructure, Estonia now faces a challenge to develop capacities and capabilities within the public sector to take advantage of the public digital infrastructure.
Thus, perhaps the most significant question faced by Estonian digital government is whether the main reasons for its success — particularly its charismatic leadership-based informal networks and civic hacker culture — provide enough capacity to harness the potential of Estonia’s digital infrastructure for more inclusive public services and society.
While the decentralised digital agendas of line ministries have provided needed agility, they have also created uneven digital capabilities across different departments and agencies. This reliance on bottom-up departmental initiatives seems to necessitate stronger and perhaps more formalised coordination structures than are currently present.
Furthermore, while e-voting is increasingly popular, other aspects of digital democracy, such as civic engagement, have remained weak (with the notable exception of the so-called Citizen Assembly of 2012–2014, which, however, failed to deliver any significant results).
Estonia’s digital success brings forth at least three lessons for the future of (digital) state:
- Economic efficiency gains are not enough as value frameworks for digital transformation. Digital agendas should be more comprehensive in focus and combine social justice and other pressing socio-political issues with economic efficiency.
- While digital infrastructure — from data registries to identification and payment systems — are sine qua non for digital governments, so are institutional innovations that would create, as John K Galbraith put it, “countervailing powers” to existing powers and routines within the bureaucracy and but also in the broader political landscape. Such examples would be public ownership option of private data.
- Public sector organisations need new forms of capabilities that centre on socially conscious design and software skills in order to harness the power digital technologies for greater common good.
Estonia is indeed in many ways Silicon Valley of digital government, in the good and in the bad. Its success is based highly decentralised and agile actors that are goal-focused but also with little regard to social outcomes.
Read the latest working paper from IIPP: Estonia’s digital transformation: Mission mystique and the hiding hand
In March 2018 I participated in a conversation on “(K)ein Traum von der digitalen Demokratie: Big Data als Chance oder Risiko?” as part of the Grenzgänger Wissenschaft talk series together with Professor Neuschwander of the HTWG Konstanz . We have talked about a number of questions and here are the answers I prepared:
What kind of data is there?
We currently distinguish four types of data:
- Administratively collected data: Surveys, or transactions in which we are open-eyed. Very detailed data that are requested and deliberately filled in. Highly structured.
- Open data: This means that some of this data is published, for example, by the public administration on open data platforms in aggregated and machine-readable format. Sometimes already helpful evaluated or visualized (unemployment figures)
- Data generated by users or citizens: Outside a formal transaction, on social media channels, other Internet transactions (shopping on Amazon), but also crowdsourcing actions, weblogs, clickstreams, search data. Can be private or public.
- Automatically generated data from human and physical sensors: measuring probes on buildings, buses, police body cameras, collected automatically without human intervention or consent. Advantage: very comprehensive data collection (all potholes through which a bus passes, complete data sets). Tracking of geographical locations, e.g. when you open your weather app.
Taken together, these are each in themselves “huge data sets” – some of a private / non-public nature, some of them public. All in all, this is unstructured Internet-generated data that is linked to structured data sets and can also contain geo-tags.
Where do we encounter Big Data in everyday life?
- Internet resources: Social media interactions, mobile phone apps, videos, photos that are shared, online search behavior, Google Nest in homes, etc.
- Structured data: Online shopping, mobile phone networks (who phones with whom), email exchanges
- Geo data: Automatic login to mobile phone poles, satellites
What do you think are the 3 biggest risks and what are the 3 biggest opportunities?
- Previously unimagined insights into the behaviour and preferences of citizens
- Quick data availability and decisions in real time (nowcasting)
- The potential of democratisation: Who will be heard?
- Distribution of Fake News
- Transparent Citizen: We do not know which algorithms are used or how they affect citizens.
- Political and economic decisions are influenced. Google Flu Trends
Where does the data actually come from? Do we make them ourselves?
Each of us is involved in the creation of Big Data every second. This happens through each of our online interactions (be it Google search, Amazon shopping, social media channel interactions, smart home, smart metering, fitness wristbands, smart phones and automatic log-in to phone poles, calls we receive, emails we send, streaming services such as Netflix or Amazon Video. Thus we leave behind so-called digital traces. Mostly passive and inactive, even without our knowledge and even if we do not actively use our devices. Do not participate to share our data. Even if you are not actively involved in a social media network, the exclusion says something.
What skills do citizens need in relation to Big Data?
There are two ways to protect yourself: First through personal actions and then also through systemic changes.
I. Rethink personal behavior:
a) Do not share everything immediately with the full power of emotions on social media! According to the latest, largest MIT study, it is clear that fake news is distributed faster and further than truths and the damage is already done. So first think about whether you want to be part of this machinery, like Pizza Gate during Hillary Clinton’s US presidential campaign.
b) Maybe read an article or the text format instead of a YouTube video or an exciting TV news show, so that you don’t let yourself be influenced by the pictures. Creating an emotional distance to the news.
c) But consider whether this can be true at all
d) Interpretation of who shares what and how and what their motives can be
e) Use offline networks such as clubs and village communities more than networking in the artificial online world.
II. Systemic changes necessary:
- Information and data literacy: articulate information needs, localise and retrieve digital data, information and content. Assessing the relevance of the source and its content. For storing, managing and organizing digital data, information and content.
- Communication and collaboration: interaction, communication and collaboration through digital technologies, taking into account cultural and generational diversity. Participation in society through public and private digital services and participatory citizenship. To manage his digital identity and reputation.
- Creation of digital content: Create and edit digital content Improve and integrate information and content into existing knowledge and understanding of the application of copyrights and licenses. Knowing how to give clear instructions to a computer system.
- Security: To protect devices, content, personal data and privacy in digital environments. Protect physical and mental health and be aware of digital technologies for social well-being and social inclusion. To be aware of the environmental impact of digital technologies and their use.
- Problem solving: Identify needs and problems and solve conceptual problems and problem situations in digital environments. Use of digital tools to innovate processes and products. To keep up with the digital evolution.
From the state’s point of view, too, the digitisation and use of big data in public administrations is important. What are examples of this?
Use of Big Data in public administration
- The first Big Data study in public administration was a combination of scientific data with Big Data from social media data (Twitter): The U.S. Geological Service was the first public administration to not only use scientific data on the magnitude of earthquakes, but also combine it with social media data to find out the impact of earthquakes on the affected citizens. By using these so-called humane sensors, it is possible to determine more quickly which decisions have to be made in the event of a natural disaster.
- To protect against terrorists: analyze large amounts of data, check for anomalies, investigate forensic evidence and help avoid terrorist attacks. This can be done with sensors on physical buildings and then synchronize the data in real time with other databases, analyze telephone traffic, bank connections, online shopping, etc..
- Use of VAT payments on online platforms are already actively analysed by tax offices in all OECD countries. All participants in economic transactions are provided with risk indicators, so that the tax office knows which transactions are risky (because they fall out of line) or which transactions are normal for a certain buyer/seller. Theoretically these analyses happen overnight and in the afternoon the tax investigator is already standing on the mat and tries to collect the allegedly evaded money.
- Use Big Data to predict the financial health of individual companies, cities or regions. In combination with various data sets, the public administration can diagnose whether a company can survive in the market and is on the verge of bankruptcy. This is important information for the public administration, as it affects jobs. This leads to increased social expenditure, such as unemployment benefit, or even the brain drain from a region, because the unemployed have to move to where jobs are available. It is therefore in the public administration’s own interest to use all available data sets to determine what is in store for them in the future.
- Government and jurisdiction should, however, be involved in the regulation of large social media companies, search engines and sales platforms:
- o hold companies accountable for allowing so-called fake news to be distributed. For example, Twitter and Facebook have only now, under pressure from hearings, looked for how Russia has placed purchased advertising in the news feeds. This means that both companies clearly benefited from this propaganda, but did nothing about it. Researchers find this propaganda very simple – so I wonder why the social media companies pretend that they have to search for it for a long time and
- “Weaponizing the Web”: YouTube as a place of radicalisation for young people and supporters of terrorist groups
- Establish ethical principles for the use of online media that social media companies must also adhere to. NetzDG (Network Enforcement Law) = Law to prevent hate speech and hate speech on the Internet. Decisions should not be made in Silicon Valley, however, but in our linguistic area, in which we understand the nuances of language (irony, sarcasm) and also the context and thus do not block wrongly criticised content.
What do you think would happen if you completely ban the collection of data?
Data collection, e.g. of Facebook data, is already prohibited in the EU, but it has become clear that the EU is not taking action and is really checking whether the data collection (the associated sale) is not actually taking place. One reason for this is that the servers are located in the USA.
From my point of view, the users are particularly in demand:
- Everyone should think about what they share on social media (parents who present their children publicly to get a few uhs and ahs)
- Do you only want to use social media personally, for example, or professionally? Add colleagues, then?
- Great discipline what you say online.
- We have become aware of some risks and threats to democracies in the context of Big Data. In the beginning, however, you also had three chances each named by Big Data for democracy. We know you look at the subject neutrally from a scientific perspective BUT what would you say if you had to make a flaming plea for Big Data?
Democratization effects of Big Data:
- There was a time when we all thought that Big Data had a democratizing effect. Everyone has pointed out that the Arab Spring would not have been possible if the demonstrators in Egypt had not gathered online worldwide to meet physically in Tahier Square. These are undreamt-of possibilities that Big Data offers: Information is distributed to many people and the power these demonstrations have can actually change the course of a government. Whether the results are always what you hoped for from a movement is another question. There are many examples, such as #BlackLivesMatter hashtag, NRA student protests in the US against the government and against the National Rifle Association -> no great improvement for the affected groups.
- On the other hand, hashtags like the #metoo campaign skipped the channel and led to accusations, job losses, publicity that could possibly lead to a change in behavior. In any case to an empowerment (strengthening) of the position of women,
The promise to gain unexpected insights into the actual behaviour and preferences of groups of people or even whole nations:
- Be it political voting behavior,
- Purchasing preferences,
- Effects of natural disasters on entire regions,
- But also reactions of citizens to changed laws and the effects on certain population groups that were not previously on the radar
Additional material is available here:
- Article “Big Data in Public Affairs” in Public Administration Review [free pdf file]
- Article “Big Data in Public Affairs Education” in Journal of Public Affairs Education (JPAE) [free pdf file]
- European Group of Public Administration keynote speech 2017: Big Data in Public Affairs (full Powerpoint presentation available on ResearchGate)
This semester I am teaching a new seminar that focuses on a relatively new topic: The challenges that public administrations are facing when they are aiming to digitize their analog services. Under the catchy term “digital transformation”, many are looking to not only move from analog to digital, also redesigning, automating, or abandoning outdated administrative acts and the corresponding services.
The term was adopted from the private sector, where digital transformation of products and business models started to occur with the use of the Internet as a distribution and communication channel. The tricky situation for public administrations however is that they can’t reinvent their business models, look for new customer segments, or abandon offline products/services. Unfortunately, public administrations are compared to those whose core business model was digital transformation of their own sectors, such as Apple’s iTunes platform, Skype replacing landline phones, Amazon transforming book/retail sector, Twitter as a newsfeed replacing traditional print newspapers, etc.
We have very limited literature on the topic, therefore the goal of this seminar is that students are deriving research questions from expert interviews and are adding necessary insights by conducting additional interviews. I will give a short introduction to the topic, plus train them in qualitative data collection and analysis – with the hope that they will be equipped to design an interview guideline, select interview subjects, analyze, and present the data.
I was able to cooperate with four external partners for this project: the City of Konstanz, the Initiative D21 (responsible for Germany’s annual e-Government monitor), the City of Ulm’s Verschwoerhaus (an innovation lab), and the Deutschen Städte- und Gemeindebund. They will join us in person or via Skype, present a short introduction of their main problems and findings and then we will open up the conversation for the student-led Q&A.
Finally, the students will design posters with their main findings – an alternative way of communicating research insights to an audience – and will discuss their findings in the Mayor’s office with civil servants who are interested in digitization.
The project is supported with an internal grant to increase the transfer between research, teaching, and practice. This is an initiative that was requested by the student body and in my opinion an applied topic like digitization fits extremely with this mission. I am also hoping that the students are gaining valuable methodological and communication skills, will be able to ‘translate’ their academic insights in plain language to a larger audience, and will be generally best prepared for their BA theses and the job market.
I have made good experiences including practitioners into my classes before, but never transferred the findings of the students from academia into practice – except for using a class blog to encourage the students to write for digital media outlets. It created a bit of press attention and was listed by FedTech magazine as one of the “50 must-read federal IT blogs” in the US. You can read about my experiences and download the syllabi here:
- Mergel, I. (2016): Big Data in Public Affairs Education, in: Journal of Public Affairs Education, 22(2), pp. 231-248.
- Mergel, I. (2012): The Public Manager 2.0: Preparing the Social Media Generation for the Networked Workplace, in: Journal of Public Affairs Education (JPAE), 18:3, pp. 467-492.
Here is a poster we designed that provides a few insights (in German):
I gave an interview to Marquis Cabrera which is now online on HuffingtonPost:
– Dr. Mergel: You’ve had a most interesting academic career; the convergence of academia and industry (and tech!) in your published works is incredible! What made you decide to commission a report on global digital services?
We currently see all kinds of organizational arrangements emerge in government: innovation or policy labs, innovation offices, and digital service teams.
My goal with this report was to understand how the bureaucracy can absorb new organizational arrangements and approaches and scale them up in government. The report focuses mostly on 18F – a digital consultancy that provides services to clients at all levels of the U.S. government. I also included a brief overview of similar international initiatives, like the UK’s Government Digital Service that served as the role model for 18F’s creation.
– In your report, you mentioned global government digital service consultancies and agencies; 18F, GDS, and DTO. To level set with our readers, what are these organizations? What factors (specific or high-level) in aggregate lead to the rise of these organization types? And, how have they had a profound influence on society? Continue reading “Interview on global digital service teams in HuffingtonPost”
Expert hearing in the German Parliament on June 21, 2017
“Modern State – Opportunities Through Digitization”
Written statement provided by Prof. Dr. Ines Mergel, University of Konstanz, Germany
1. In the process of digitization, it is important that state and administration modernize their exercise of functions and fully utilize the opportunities of digitization. What, in your view, is the present state of administrative modernization and where is the potential of digitization being used in what manner? What are the success stories in Germany? What past successes can we build on?
The level of administrative modernization and digitization of the public administration in Germany is continuously declining in recent years. While legislative measures, such as the Digital Agenda, privacy policies, or investments in broadband services are being advanced, it is difficult for the German administration to improve in the e-government rankings. In a global comparison, Germany was placed at number 17 in 2011 and fell in the United Nations World e-Government Ranking four places to No. 21 in 2014. Compared to the rest of Europe, Germany occupies the 20th place in the field of digital service offerings according to the 2017 DESI ranking. In comparison, the leading e-government countries, such as Estonia and Denmark, began their digital transformation of the public administration 10 years ago.
Reasons for this are manifold. A McKinsey study to “E-Government in Germany – a Citizens’ Perspective” from 2015 shows, that the use of existing e-government services has stagnated since public administration digital services are not user friendly from a citizen’s perspective. According to the 2016 DESI study, only 19% of Germans use the online offerings of the public administration. This means that investment in e-government services fizzle out and bring no added value for citizens.
I wrote a paper based on my interviews with CTOs and digital service innovators in the U.S. federal government. The goal of the paper is to bring together the elements that lead to innovations in digital service delivery. I contrast traditional software development processes with elements of an agile innovation management approach. The result is a research framework and research questions for future explorations:
AbstractGovernments are facing an information technology upgrade and legacy problem: outdated systems and acquisition processes are resulting in high-risk technology projects that are either over budget or behind schedule. Recent catastrophic technology failures, such as the failed launch of the politically contested online marketplace Healthcare.gov in the U.S. were attributed to an over-reliance on external technology contractors and failures to manage large-scale technology contracts in government. As a response, agile software development and modular acquisition approaches, new independent organizational units equipped with fast reacting teams, in combination with a series of policy changes are developed to address the need to innovate digital service delivery in government. This article uses a process tracing approach, as well as initial qualitative interviews with a subset of executives and agency-level digital services members to provide an overview of the existing policies and implementation approaches toward an agile innovation management approach. The article then provides a research framework including research questions that provide guidance for future research on the managerial implementation considerations necessary to scale up the initial efforts and move toward a collaborative and agile innovation management approach in government.