Policy and technology: a "longue durée" view

Random thoughts on policy for technology and on technology for policy



Three innovations in digital government KPI

As practitioners know, progress in digital government is painfully slow. One can see this in key events such as EU ministerial conferences, where the debate seemed eternally repeating itself.

Yet something has changed over the last years. I saw that in Tallinn in 2017, and recently at the Helsinki ministerial conference. Progress is more real, the debate is more concrete. The gap between EU policy conferences and real-world has narrowed.

One sign of this progress is the KPI mentioned in the official speeches to show progress and success.

Back in 2001, at the time of the first conference in Brussels, the KPI was how many million Euros were going to be spent on projects.

At Manchester in 2005, the usage of the eGovernment indicators had become widespread. The KPI was service availability: what % of services were fully interactive. This was a stunning success of the “benchmarking eEurope” indicators first established in 2001.

So this indicator remained used and abused for many years. Around 2008, at the time of the Malmoe conference and declaration, another indicator started to be frequently mentioned: the uptake by citizens and business, measured through Eurostat surveys. This showed a permanent lag in citizens adoption of eGovernment, and the limitations and perils of a supply side approach.

In Tallinn 2017, these percentages were complemented by data on the number of transactions, as produced directly by the services. Member states established dashboard showing how many payments citizens made or how many certificates they downloaded. The indicator is similar to the previous one produced by Eurostat, but instead of asking citizens whether they used eGovernment, governments start using the data they already had. In other words, by reusing the data held in their databases governments applied the once-only principle to benchmarking. 

And last week, in Helsinki , a new KPI was used, measuring access data to base registries and traffic through data exchanges. Government used the data they hold to show not only service uptake by citizens, but progress in back office reorganisation and adoption of base registries, by other public administrations and also by private companies.

This is a remarkable evolution. We are getting real. Indicators have evolved in three positive directions:

  1. From input to output to adoption.
  2. From front-office to back-office.
  3. From surveys to reuse of administrative data.

It might be the effect of the adrenaline from the conference, but I am much more optimistic on European Digital Government. And even if it is just adrenaline – this is a remarkable result for a digital government conference.

Why large companies are more productive, but large states are less effective?

At our recent event on ICT and productivity, both speakers showed how large companies are more productive than small companies – because they adopt more technologies, have more skilled workers, train their workers more.

screenshot 2019-01-07 at 20.52.54

Why then, when it comes to government effectiveness, the opposite is true – small states are more effective than large ones?

Here is a chart from the world bank indicators on government effectiveness.

screenshot 2019-01-07 at 21.34.15



A paradox: increasing support for globalization in the age of populism

Screen Shot 2018-03-06 at 18.01.55.png

I was surprised to discover today that support for globalization has been steadily growing in Europe over the last 6 years. In 2017, 62% of European consider it a good thing, up from 47% in 2011. Source: Standard Eurobarometer 85, spring 2016.


An Open Letter to @ArlingtonVA School Board

As a parent of two kids attending a Montessori school in Arlington, VA, I am deeply worried by the proposed budget that removes all teaching assistants from Montessori Elementary Schools.

The Montessori method requires the presence of skilled teachers. You can’t make education more efficient by reducing personnel. Especially in Montessori, there is no room for economies of scale. The thing about Montessori is that it has to be done very well to be effective, you can’t do it halfway. And when it is done well, the results are amazing.

We are from Barcelona, in Europe. We care for our schools. We sent our kids to the best schools we could find in Barcelona, we spend ages choosing the right school. We are proud of our school in Barcelona. But boy, we found nothing comparable to the Montessori school we found here in Arlington. Our kids, now, are just craving for learning. They don’t want to watch videos or play video games, they want to paint with Metal Inset and write stories.

We are living difficult times, tensions and conflicts. Our society is becoming more and more divided. We can wonder about the big theories of inequality, but the fact is that the main factor explaining inequality is education. You have a great, world-class education here, with the Montessori school. This is a gift, a privilege. You should protect it in any possible way.

And by the way, we didn’t expect this. We thought education was going to be a problem here, that good schools in the US were only for millionaires. Yet, admittedly with some sacrifice, we could afford the best school we ever found.

So, I want to conclude with one word: thanks. To Arlington County, to our school, to our teachers. Thanks for giving this opportunity to our kids. I hope that I will be in a position to continue thanking you in the future.


Why the Free Flow of Data matters for SMEs

Data is the new oil. But the peculiarity with data is that by itself, it’s not worth anything. To create value, companies need to have not only the data, but the skills and the “data mindset“. And many large companies in traditional sectors typically have huge amounts of data, but lack the skills and the mindset. To extract value from these data large traditional companies have to collaborate with small data analytics companies that are able to transform data into information and knowledge. This is why BASF worked together with Yukon Digital (a data analytics SME) in order to deliver predictive maintenance for their chemical industrial plants.

But we don’ have to think of SMEs as mainly data analytics companies: in their vast majority, they are traditional businesses which typically lack the skills, and mindset, and often the data themselves. Yet even in this case, they can benefit from big data solutions by collaborating and sharing data with other data-driven companies. What often happens, in this case, is that SMEs can partner with companies that actually gather the data through sensors.

For instance, small transport companies can optimize their fleet management by installing devices in their trucks, such as those provided by . The supplier takes care of providing the devices, gathering the data and analyzing them. Shop owners can use “beacon apps” to understand how customer move and interact in the shop. Farmers use sensor-equipped tractors that include analytics services.

Yet, despite these opportunities, both large and small companies are typically reluctant to share their data. Even for non-personal data, which do not have privacy implications, it is not clear who owns and control it, and there is limited transparency over what different companies can do with it. SMEs often lack the skills not only to make the best use of their data, but to understand the terms and conditions that govern data-driven services.

This is why the European initiative on the Free Flow of Data is so important. Data sharing across company and country borders brings benefits to SMEs – whether they are big data startups or traditional business. But there is a need to ensure trust and the right incentives for companies to share data. It is not about establishing rigid rules about who owns the data – but about creating the conditions for the widest reuse of data.

Because, after all…

The best thing to do with your data will be thought of by someone else

A hard look at European e-government: have we made enough progress?

Eight years are an era, in Internet time. And this is the time passed since the last e-government Ministerial Declaration, signed in Malmö in 2009. To put it in perspective, at that time Nokia and Blackberry had 70% of the market share in mobile.

What has changed since then in European e-government? A lot of progress has certainly been achieved – but is it enough?

On the bright side, many things have happened. The revised PSI directive introduced the “open data by default” principle; the eIDAS regulation set the standard for mutual recognition of ID across Europe; and the new European Interoperability Framework presents both a modern vision and an implementation roadmap. The “building blocks” architectural approach has been widely embraced and composable services such as eInvoicing, eDelivery, eSignature, eID, and eTranslation are already being deployed. Governments have deployed online services , and have started to bring in new skilled workers through “Digital Service Teams“.

Yet the results cannot be considered satisfactory if we take as a yardstick the objective of the Malmö declaration [pdf] of “services designed around user needs” and ·”seamless eGovernment services for the setting up and running of a business and for studying, working, residing and retiring anywhere in the European Union.”

Let us just take one example. Before Malmo, in 2008, 32 of Europeans out of 100 shopped online and 17 used government services – a 15 points gap. Today, 55 shop online and 28 use government services – a 27 points gap. Europeans are today accustomed to “shop online first” but still consider using public services online as an exception. The gap has widened, and any person having used online public services knows that in general the usability leaves much room for improvement compared to private services. The only existing European seamless services are private.

Figure: Uptake of e-commerce and e-government in EU (Source: Eurostat)


To be clear, this is not the only problem – but a sign that e-government is not (yet) having the desired impact. And in our historical context marked by continuous decline in trust in government and the rise of populistic movements that threaten free trade, democracy and ultimately peace, we simply cannot afford it.

So what are the underlying reasons for this limited impact? We see different progress with regard to design, implementation and deployment.

In terms of design, the progress is remarkable. We moved from an obsession towards online service delivery to a more balanced approach between front and back office. The strategic approach has moved from monolithic systems to open, composable building blocks. The vision set out in Malmö and deployed with the eGovernment Action Plan remains valid and it is now accompanied by the European Interoperability Framework 2.0 – an impressive document worth reading.

In terms of implementation , the progress is patchy. The building blocks are being delivered successfully. Implementation of PSI directive is practically complete by all MS, while for eIDAS it’s still at an early stage. The EIF implementation is still partial. Overall, there are some excellent, truly end-to-end services and some very advanced countries, but the landscape is very diverse and unequal.

In terms of deployment, the progress is insufficient. Governments show a limited capacity for engagement of key stakeholders across the value chain: other administration levels implementing interoperability guidelines, stewards of base registries publishing data in standard format, private service providers using building blocks, developers reusing open government data, citizens using public services. What we need is a truly end-to-end digital government experience, that can only be achieved by involving all relevant parties from the early design phase.

We need to bring all stakeholders together under a common vision, and infuse renewed energy into innovation in government. This is why we need a new EU ministerial declaration.

On behalf of the Estonian government, the Lisbon Council has presented some initial proposals for discussion.

Now it’s your time to speak. Do you agree? What is missing?

Comment our proposals, let European Member State know what you think.

We all need a better digital government – let us know how we get it.

We need ethical principles for news sharing, not just for journalism


I keep receiving fake news, the latest in Whatsapp. What can we do about it?

There has been a shift in journalism. Media are no longer the gatekeepers: people get and share news on Facebook and alike.

This is a good thing: it’s a democratisation of news sharing. But it doesn’t work now because people are not educated about what and how to share. They struggle to distinguish truth from falsity. As Carlota Perez would say, we have the technology, but we don’t have the institutions.

This is not strange. This is normal. You have innovation, then you set up norms that help making the best of it.

Some people suggest that these platforms should monitor false news. This is one form of governance – moving the responsibility from mass media to social media. I don’t like it, I think it is paternalistic and I certainly do not trust Facebook more than traditional media.

I think education and responsibility is more important. Journalist have ethics and principles: if people start taking over the job of journalist, they should learn from those ethics.

We need to jot down a set of principles for news sharing:

  • Read the message fully and carefully. If you feel an immediate impulse to share even without reading, be even more careful: it is probably because the message is deliberately designed to push you to share.
  • Verify the source of what you share. Don’t share anything that you can’t trace the origin of. Even better if you have two different sources, and perhaps reputable ones. It’s not difficult.
  • Verify that the message is not already debunked as a scam. Just google it.
  • Be suspicious of any message that openly ask you to disseminate it. They are most likely scams.
  • Do not copy paste messages, especially if they contain first person verbs. I just received a message in different Whatsapp groups saying “I have friends in the police”. Of course I trust a message if a friend of mine says he has direct insight, but the truth is he did not have these friends, he was copy-pasting. If you copy-paste, you are directly lying to your friends.
  • And most of all, remember the tale of The boy who cried Wolf! If you share false news, people won’t listen to you.

What other principles can we add? And can we make a kind of self-certification for these principles to show, for instance, in our Facebook photo profile?

If Trump is a zig, how long before we zag?

“The path that this country has taken has never been a straight line, we zig and we zag.”

This is what Obama said to reassure americans about the Trump presidency. Trump is considered just a phase, an antithesis in Hegelian terms. It is temporary. It won’t last. People will quickly realize their promises won’t be fulfilled and will return to sanity.

I beg differ. What we see is different. Populist leaders, from Putin to Erdogan to Berlusconi to Netanyahu, they are not fast to go even if clearly they do not deliver.

Take Berlusconi: when he was elected, we thought it was just a moment of craziness. Perhaps the most authoritative Italian journalist, Indro Montanelli, famously declared in 2001 that Berlusconi was a plague that could only be cured by vaccination, i.e. by placing him into governments. Then Italians would realize his promises were unfounded, and they would vote him out the office. It didn’t happen, at least for 10 more years.

Populists are masters in finding scapegoat, external enemies to explain their lack of success. Berlusconi blamed the rigid political system, the press, Europe, the judges.

And he got away with that.

Because public policy is difficult, success is hard to define, especially in the era of post-truth.

I don’t have a solution, but I know that populism doesn’t go away quickly, easily or by itself.


You don’t have access to real-life big data? Just create them through simulation!

As I previously wrote, lack of  access and reuse to corporate data can be a bottleneck for developing Artificial Intelligence.

One alternative is to re-create the data through advanced simulation. Simulation helps creating massive quantity of data from the so-called “digital twins” of industrial “cyber physical systems”. Obviously simulated data are a simplification so they might miss several real-life issues, but at the same time you are able to experiment with a much wider set of potential situations that are not yet available in real-life – such as a disastrous failure in a chemical plant.

Are simulated data good enough to substitute real sensor-generated data?


Blog at

Up ↑