In “Open Government Data,” What’s Really Open?

In the fall, the United States joined seven other countries — with hopefully more to come — in forming the Open Government Partnership, an international initiative designed to hold countries on a global stage to a commitment to open government and its close relative, “open government data.” These terms, though, have created a bit of philosophical and grammatical confusion.

“Eventually it became clear to us that it was really the word ‘open’ that caused problems,” said Harlan Yu, a doctoral candidate in computer science at Princeton University, who has written a new working paper on the topic with Yale law student David G. Robinson.

“‘Open’ can mean so many different things. Especially with a term like ‘open government data,’ it’s really not clear whether the adjective ‘open’ modifies the noun ‘government data.’”

Or is it the other way around? Does “open government,” in the traditional context of transparency and accountability, modify “data”? Put another way, are we talking about real-time transit information (data collected by a government and freely given to the public), or White House visitor logs (data about government itself that creates true transparency)?

Yu and Robinson argue that these two meanings have blurred. It’s increasingly unclear just what people are referring to — and promising — when they say “open government.”

“You could see an argument in favor of it in principle, this merger of these two ideas,” Robinson said. “If ‘open government’ is a bigger tent that encompasses a lot of these different beneficial things, it’s easier for people to get behind it. But the price you pay for that, we would argue — the price that is inherent in that merger — is that it gets harder to talk about the traditional open government stuff.”

That is, the stuff of accountability and transparency. Amid this confusion, governments with no real designs on transparency may don the mantle of “open government” simply by publishing, say, weather data online. Yu and Robinson cite the Hungarian city of Budapest, which releases web-based, machine-readable public-transit schedules that enable apps like Google Maps to plot route information. This undoubtedly improves life for anyone who lives in Budapest or passes through it. But Hungary hardly has an open government in the traditional sense. It may, in fact, be reverting to authoritarianism.

Yu and Robinson argue that we’ve begun to confuse the politics of open government with the technologies of open data. Technology is making that real-time transit data more accessible than it’s ever been. But open government is not really about making information that was always public easier to access. It’s about making information that was never previously open — those White House visitor logs, for instance — public for the first time.

Click here to view larger.

This was the original meaning of open government. In an attempt to finger exactly when the terminology shifted, Yu and Robinson trace the history of the concept back to World War II. The American Society of Newspaper Editors commissioned a report in 1953 that includes what may be the first mention of “the public stake in open government.” Their advocacy led in 1966 to the federal Freedom of Information Act, and for the next several decades the idea of “open government” referred unambiguously to the disclosure of previously locked-up information.

The rise of technology started to change this in the 1990s. Suddenly, plenty of things that had nothing to do with government were “opening” up: open-source software, open-access scientific journals, Creative Commons copyright licenses. The federal government embraced this in 1995 when it created THOMAS, the web portal that tracks legislation in Congress. That information, though, was never secret. It was just hard to get.

More recently, the Obama administration has broadly embraced both open government and open data, but it has also, the authors suggest, done much to confuse the two. Plenty of recent initiatives reflect both principles. But it’s also possible to embrace one without the other, to release, for example, new lobbying data in a format watchdog groups can’t easily sort, or old government mapping data that has nothing to do with accountability.

“Data that the government itself already collects is ‘government data’ because it’s government-collected,” Yu said. “But a lot of this data which is really useful for citizens isn’t necessarily about the inner workings of the government itself. It’s not necessarily going to expose corruption or make government processes become more accountable.”

Yu and Robinson propose clearing this up by assessing open government and open data initiatives along two related axes: a question of technology — how easy is the data for outside groups to reuse? And a question of transparency — to what extent does it help produce public accountability? Using this more nuanced framework, ideas can be plotted on a two-dimensional map that acknowledges there’s such a thing as open data that doesn’t open governments, and open-government revelations that don’t come in easily used databases.

The problem Yu and Robinson are trying to solve is a welcome one (even as it points toward the prospect of some nefarious government behavior). These terms have become confused largely because so many people from different spheres now want to use them. And that’s a good thing.

“It’s a great problem to have, and we’re thrilled to find ourselves here, given where we were four years ago, or 10 years ago,” Robinson said. “This is about a midcourse adjustment. But fundamentally, we’re steering in roughly some good directions, and we’ve got a significant amount of momentum.”