Transparency Is Not An End In Itself

May 13, 2013 Dennis D. McDonald

Much is being made of opportunities that might be opened up by the President’s recent executive order making open and machine readable the new default for government information. It’s not all about policy goals like transparency and public participation, it’s also about innovation and jobs. For example, on a recent trip President Obama said,

“So over at the Capital Factory, I met with folks behind the start-up called StormPulse, which uses government data on weather to help businesses anticipate disruptions in service. And then you’ve got a Virginia company called OPower that’s used government data on trends in energy use to save its customers $200 million on their energy bills. There’s an app called iTriage, founded by a pair of ER doctors that uses data from the Department of Health and Human Services to help users understand medical symptoms and find local doctors and health care providers. And today I’m announcing that we’re making even more government data available, and we’re making it easier for people to find and to use. And that’s going to help launch more start-ups. It’s going to help launch more businesses.”

Some cautionary notes have been raised by sources such as Ars Technica where Sean Gallagher discussed the implications of the new executive order:

Just how effective this order will be in the face of the government’s ongoing budget crisis is unclear. With sequestration cutting back many programs, there’s little maneuvering room for agencies to make significant changes to the systems that this order would affect.

I come down more on the cautionary side. While I have a fundamental belief in making government data more accessible to the public — I’m a citizen too and want to know how my tax dollars are being spent — I’m also thinking beyond accessibility and transparency to consider whether and how “open data” will actually make government programs more effective.

While data generated by a government program should be readily available and shareable among its target constituencies and members of the public, I also believe that resources devoted to improving the transparency of a government program should also promote the accomplishment of that program’s goals.

We need to understand the resources involved in making government programs transparent, especially when we consider “big data” and the new and innovative uses to be made of data generated by government programs. As I’ve noted elsewhere, transparency isn’t free. Someone has to pay for setting up and maintaining the web pages, databases, and services that are needed to make government program data accessible and usable. “Big” or not, the data have to exist before they can be sliced, diced, mashed up, analyzed, shared, and/or resold. For example, if program spending data need to be standardized before they can be sliced and diced analytically, someone needs to take responsibility for standardization efforts and for the resources necessary for storing, managing, updating, and analyzing the data.

Hopefully the policies and programs to make data accessible that are currently under development will not just be focused on making data accessible but will also seriously focus on how making the data accessible will promote program goals.

While we’re at it, I also think we need to make the process by which we make data more transparent, well, more transparent. How decisions are made about which data to publish and in what format should also be subjected to scrutiny, otherwise we might find that technical features are driving accessibility and availability rather than value or utility. That might seem counterintuitive, but in times of budget issues decisions might get made based on the availability of “low hanging fruit” rather than what is actually useful.

If this sounds like I’m recommending that data access design issues be at least partly driven by user requirements, that is the case. I’m not saying that we can anticipate every possible use of the data associated with a government program; there are too many arguments counter to that. I am saying that a basic principle of system design is to take user requirements into account. The user requirements I’m talking about are the intended beneficiaries of the government program. And, if a major requirement is that “open data” stimulate innovation and job creation, we need to be able to measure the success of that as well.