What Estonia Knows That We Don't

In an October interview with Bruce McCabe, Ott Velsberg, Estonia's Government Chief Data Officer, described the next phase of his country's digital state as proactive rather than reactive — services that arrive at the citizen instead of waiting to be requested. The example he uses is early retirement. The state already knows when someone qualifies. The legal infrastructure to grant the benefit exists. The data is on file. In the Estonian model, the system identifies the eligible citizen, prepares the paperwork, and acts. The citizen does not apply. The citizen receives.

Joonas Heiter, who runs Estonia's State Information System Authority, puts the same idea more simply: "Citizens and entrepreneurs don't think they have to declare something or inquire about something. If they have some need, the state should be proactive."

This is a sentence that should embarrass anyone who has worked in American digital service delivery. It describes an architecture the U.S. state is technically capable of building, has the data to power, and chooses not to construct.

I wrote in the last piece about how the abundance debate has produced no theory of the entitlement layer — the stratum of public services where the program already exists, the eligibility is already statutory, and the household still doesn't get the thing. The Estonian answer to this problem is to remove the asking. The U.S. answer is to keep asking, every cycle, in a different form, with a different deadline, through a different agency, and to call the resulting non-receipt "low take-up."

Velsberg is careful about how he frames the work. He told GovInsider that the principle is "not to duplicate what vendors already offer" — Estonia uses Microsoft 365 like everyone else, accepts the risk that comes with that, and concentrates its own building on the layers no vendor will deliver: the interoperability spine, the data exchange protocol, the identity layer. The X-Road infrastructure that connects Estonian government databases is twenty-five years old and remains the thing American digital service efforts have never built and apparently cannot build. We have spent twenty years digitizing forms. They spent twenty years digitizing the relationships between the agencies that own the data the forms ask about. That difference is the entire ballgame.

There is a thing American civic tech tells itself about why this gap exists, which is some version of "Estonia is small, the U.S. is federalist, the cultures are different, the legal traditions don't transfer." All of that is true and none of it is the actual answer. The actual answer is that a proactive state requires the agency to do work the citizen currently does, and the agency does not want to do that work, and no political constituency exists that can force the agency to do it. Administrative burden is sticky in the United States because someone benefits from it being sticky. The vendor benefits because the form is the product. The agency benefits because every application denied is a budget line that doesn't pay out. The legislature benefits because non-take-up makes a program look cheaper than it is. Nobody at the household level has the time, money, or organizational power to change this.

The Estonian system is interesting because it shows the technical capacity is real. The X-Road exists. Bürokratt, the network of voice-driven virtual assistants Velsberg's team is building, is a real piece of working software, not vaporware, and they are exporting the architecture to other governments. The U.S. doesn't need to copy it — federation makes that hard, and the data privacy regime is structurally different — but the U.S. could absolutely build the version of it that fits the American institutional context. Code for America has done pieces of this. Nava has done pieces of this. 18F did pieces of this. None of it has been knit into the kind of spine that would let an agency proactively grant a benefit it already knows the citizen qualifies for. The work hasn't happened because the work hasn't been demanded.

Velsberg makes one more point that's worth dwelling on. He told McCabe that Estonia's approach to proactive services has run into a question they're still working through: how proactive should the government actually be? The example he uses is automatic license renewal — uncontroversial, easy, citizens like it. Then he moves to re-education programs and automated career planning, where the proactive state starts to feel paternalistic, and citizens push back. The Estonians have a phrase for this they keep returning to: "should the government become a mother figure taking care of citizens?"

This is the question the U.S. has never asked because the U.S. has never gotten close enough to the proactive state for the question to come up. American digital government is so far from honoring its own existing obligations that the worry about over-reaching has no purchase. We have entire categories of citizens who qualify for benefits the agency knows about, with paperwork the agency already has, and the agency declines to act. Worrying about the paternalism risk of proactive services in this context is like worrying about the etiquette of the dinner table while standing in the parking lot.

The thing to take from Estonia is not the technology. The X-Road is interesting; Bürokratt is interesting; the data steward model is interesting. None of it is the point. The point is that a state can decide that the friction between the citizen and the benefit they're owed is the state's problem to solve, not the citizen's problem to navigate. Estonia decided that twenty-five years ago and has been building the infrastructure to honor the decision since. The U.S. has not made that decision. The U.S. has made the opposite decision, repeatedly, in every administration, and called it customer-centric design every time.

There is a version of this work happening in the U.S. — the state-capacity wing, the Recoding America crowd, the post-18F practitioners building under-the-radar inside specific agencies — but it operates without a forcing function and without the political mandate Velsberg's team has. They're trying to retrofit a proactive layer onto a state that was never designed to be proactive, against an institutional culture that treats every application as an opportunity to deny.

Velsberg is in his second decade of doing this work. The U.S. has not started.