I was absolutely delighted to present at the Gov2.0 Summit here in Washington this morning. Thanks to Tim OReilly and his whole crew for the invitation and all the hard work they put into the event.
I have to admit it was a real challenge to present an "Open Government Scorecard" in the 10 minutes I had allotted to me but as I prepared my remarks I found that the time constraint really made me focus. There were a number of disappointments that I have in the President's record that I didn't mention, such as his calling for -- but not acting upon -- a centralized database for earmarks, or his promises to push for lobbying reform that has never materialized -- but I hit some of the major concerns that I and my colleagues at Sunlight have with the Open Government Directive, Recovery.gov, Data.gov, and USA Spending.gov. (Tomorrow I'll talk more about our ongoing analysis of the data on that site.)
And rather that summarize my remarks, I'm presenting them in full below. Let me know what you think in the comment section below.
From the Executive Memorandum on transparency issued by President Obama on the first full day in office, to the full-blown Open Government Directive, to establishing new policies regulating lobbying of the Executive Branch, no organization has been more excited, enthusiastic or optimistic about the advent of Gov 2.0 for data transparency than Sunlight. In many respects, this Administration has gone further and faster toward creating a transparent government than any that’s preceded it.
But now, 20 months later, the drive for transparency appears stalled.
The Open Government Directive -- the Administration’s 10-month-old manifesto on government transparency -- can either become a dated relic destined for the ashbin of history, or a transformative commitment to a new era of openness. Right now it’s teetering someplace in between.
As you may recall, the central thrust of the directive is its insistence that all government departments create and implement their own open data plans, aimed at releasing “high-value” data to the public. The plans that resulted, however, were little more than aspirational. In the first of those plans, 12 out of 30 agencies didn’t identify any data for future publication and altogether only 75 new data sets were promised. 75 data sets? That was hugely disappointing.
Enforcement of these plans has always been ‘soft.’ And now even faces greater uncertainty with the departure of OMB Director Peter Orszag and WH Ethics Counsel Norm Eisen. The Directive is only as strong as its enforcement.
To be sure, there have been some meaningful first steps from agencies and the White House. HHS and NASA both immediately spring to mind as aggressively pursuing new projects that are on the cutting edge. The White House itself has taken some meaningful steps in posting its staffers’ ethics filings online, requiring extensive stimulus lobbying disclosure, and posting the Visitor Logs online for the first time. But these aren’t well-established policies, and exemptions to publishing this data are unclear and unstated. All of these initiatives need a steady hand and a clear commitment to mature into permanent, reliable, effective policies.
And one central Obama campaign promise, to “create a centralized Internet database of lobbying reports, ethics records, and campaign finance filings” isn’t even on the drawing boards.
Data.gov started with enormous promise. A single catalog for all government data is very exciting in concept. As it has evolved, we have gotten a progressively better website: a more attractive design, and somewhat improved organization of the data sets. But it's still a pretty mediocre data repository and the types of data available remains an enormous concern.
It turns out that the government has some interesting ideas about what counts as "high value" information. The Department of the Interior seems to feel that population counts of wild horses and burros are "high value" but records of safety violations like the ones that seem to have led to the Upper Branch Mine disaster are not.
We want to see data that can be used to hold government – and the entities that report to it – accountable: records and data that would allow the public to assess the effectiveness and efficiency of federal programs, policies and initiatives; the competence and integrity of its employees and contractors; its management of public resources.
Recovery.gov, the government’s signature project to demonstrate transparency around economic stimulus spending is hard consider it more than a qualified success. And the idea of collecting job creation data at the most local level was overly optimistic from the start – and ended with Vice President Biden making excuses for the data’s poor quality on the Daily Show.
Now, let’s take a look at USASpending.gov, the project that then-Senator Obama and Senator Tom Coburn championed into law back in 2007.
USASpending was created to provide the public with information about how the federal government spends our tax dollars. It was launched in late 2007, but it's already gone through three redesigns, each one flashier than the next, with gradients and maps and now a sort of GapMinder-style visualization tool. It's pretty impressive…. looking. Unfortunately, its data is almost completely useless….
To demonstrate just how flawed, Sunlight is announcing today a new project today called Clear Spending.
This project – the result of the hard work of Sunlight’s Kaitlin Lee and Kevin Webb -- is an ongoing analysis of the grants data in USASpending. We started with a methodology for evaluating data quality that has been used by the GAO.
It’s really pretty simple.
Federal spending gets reported in two places. Individual transactions are sent to USASpending. There’s also a separate, somewhat redundant process where total spending for each program is reported to the Catalog of Federal Domestic Assistance, or CFDA. It’s possible to tally up the USASpending records by program, and compare those totals to the ones in the CFDA. When the two don’t match, we know there’s a problem.
That’s the heart of the method. We also looked at whether the reporting occurred on time, and whether all of the reports’ mandatory fields were filled out. Then we automated those three tests, so that we could run them against all ten million rows of the USASpending database.
The results aren’t pretty. We found just under* 1.3 trillion dollars of broken reporting in 2009 alone. That's fully half of the spending for that year. Some of the numbers are too big, some are too small, some are missing completely. Others don't have the detail that's required, or were reported months later than the law demands.
The data powering USASpending is broken. You can't trust any aggregate numbers you get from the site -- answers to questions about federal spending that rise above the micro level. When we say things just don’t add up, we mean it.
The government has known about this problem, and they say they're working to fix it. But what we've actually gotten is a series of redesigned websites, each one with data just as unreliable as the one before it.
We are beginning to worry that the Administration is more interested in style than substance.
If we settle for a superficial kind of approach, Gov 2.0 will be remembered as a failure. Government has learned to say the right things -- now we need government to actually get serious about technology and transparency.
There's a lot of work left to do. These are not easy tasks, or certainly not glamorous ones. But these are the types of challenges that we must undertake if the promise of Gov 2.0 is going to be realized.
For starters, we have to take on some of the responsibility for making this happen ourselves – I mean ‘us’ as in the community of Americans who are concerned about accountability. And that’s a lot of us. And we wear all kinds of political labels from conservative to liberal, to libertarian to the labels of Tea and Coffee Party activists.
Our job is to hold the Administration’s feet to the fire – bureaucrats aren’t going to act just because someone asks nicely. Government isn’t going to change how and when it makes data available – even when a few good people on the inside want it to – because of a directive.
It’s not going to happen until laws are changed, or Executive Orders are issued, or until enforcers are given real power and the President himself makes it a priority.
And finally, we need to admit that Gov2.0 isn’t happening until citizens are truly actively engaged in helping to demand and co-create it.
We can still have a government that is truly open, participatory and collaborative. But it won’t happen until we push for it.
*We updated this number - at launch it was just north of $1.3 trillion. You can see why here.