This isn’t about research or methods, so I’ll be brief.
- Cass Sunstein, US policy veteran and eminent scholar, recently released a draft article distinguishing between input and output transparency, suggesting that arguments are weaker for the former, and offering reasons why input transparency might often not be a good thing. (To grossly oversimplify: there are too many inputs to policy-making processes, and making inputs transparent is potentially costly, not very useful, and could have chilling effects on the kind of open conversation that leads to good policy).
- This was picked up in popular political discourse viz-a-viz the Clinton private email server scandal. Vox’s Matt Yglesias gives a good representation of the argument for why some political communication should not be transparent, and @ gives a good nuts and bolts argument for why government emails should be public. Sunlight Foundation puts the same pro transparency argument in the context of current FOIA amendment discussions (here and here) update: and in the context of US cities.
- Tom Steinberg revisits Sunstein’s article to reflect on how the transparency and accountability field needs to improve our conceptual and rhetorical approach to what kind of information we demand (you can always count on him for a smart field comment). His summary of Sunstein’s article is also solid, if you want more detail than the above, but don’t want to dig into the full text. His most salient point is that lobbying in the US is too big an issue to let it be exempted from input transparency.
I’m mostly interested in Steinberg and Sunstein’s take here, because I think they’re both right to assert that moving our conceptual thinking ahead will guide actual policy and advocacy discussions about government transparency, and that the transparency camp doesn’t have a good way to think about this kind of thing.
First, a kind of obvious, but knit picky point: a lot of the resistance to “input transparency” notes how politicians will use phone or face-to-face meets to avoid the transparency liability, as if this is evidence of input transparency’s negative impact. But backchannels have always been a mode of political operation (as Yglesias notes), and it’s a kind of necessary steam valve for functional governance. It’s also one of the best informal indicators we have for assessing the degree to which governance is transparent. This practice won’t ever disappear if emails are protected from public view, and its exisitence is not in itself an argument for or against anything. That behavior over time adapts to regulatory frameworks is rather obvious, but aslo implies a dynamism in transparency policy, which is often forgotten in these types of discussions.
But back to the main issue. As both Sunstein and Steinberg note, this conceptual distinction isn’t a policy recommendation, but an exploration into principled balancing. Neither propose what the next steps might look like, and seem content to address this first foray into the conceptual mess. At the end of the day, however, it seems reasonable to expect that more nuanced regulation will require a case by case evaluation of different types of information. Which officials are relevant? When is meta data about communications as important as content? Is tiered access ever appropriate? These types of questions are the ones that will likely shed light on how to deal with Steinberg’s concerns about the lobbying industry.
When we get into these kinds of weeds, Sunstein’s welfarist perspective is especially important. It’s absolutely the right approach to consider regulatory approaches from the perspective of societal costs and benefits, rather than absolute principles. But the categorical frame isn’t likely to be helpful, first and foremost because the inevitable detail required to accommodate heterogeneous cases will make it too difficult to apply.
A jurisprudential-style criteria or set of limitations and exceptions would be preferable, with a presumption of transparency for certain types of communications. For example, international human rights law allows for exeptions to free speech using a three part test: codification (is it spelled out in law), substance (does it have a legitimate aim, like protecting public order or national security), and necessity (is there really no other way to protect that aim) (more here). Of course I’m not suggesting the substance of that particular test, but the modality of presuming transparency and then addressing exceptions from different perspectives might be a smart way to balance the conflicting norms at play here.
Most fundamentally, though, I want to note that the dichotomy itself is irksome. The categorization of input and output transparency is essentially about policy-making processes, which is only part of governance. Not making this explicit risks subsuming all the other kinds of government info that should be considered for transparency under input transparency, and subject to a more conservative standard. Things like information on government procurement, politician affiliations,or archive maps for government data holdings can easily be thought of as policy inputs if inputs and outputs are the only options, but they’re qualitatively different, and fundamentally important for the conditions of open and transparent government.
With that in mind, I agree that we need better conceptual and rhetorical tools for thinking about when government communications should and shouldn’t be transparent, but this dichotomy is too simple a place to start.
Well, so much for brief.