 So this is work very much in progress, so thoughts are really welcome. So I'm going to be defending with every bone in my body output transparency and raising some questions about input transparency. These will be defined. I'm going to be suggesting a well-fairest approach to the Freedom of Information Act and to Freedom of Information generally, suggesting kind of block-headedly that we should be focusing on the human consequences of transparency policies. What does it mean for real people who are often struggling with environmental risks or poverty or highway safety problems or joblessness. How is transparency affecting them and focusing much less on the word legitimacy or in the notion of transparency as important for its own sake. So that I'm not going to embrace. The structure is going to be simple and in four parts. I'm going to say a few things about the Obama administration. Some things about outputs and some things about inputs and some things about the future. Okay, so a Marja Sen, Harvard's own, has an empirical finding, a stunning one. No famine in the history of the world has ever experienced a famine in a functioning democracy. That's stunning because you'd think wouldn't you that a famine is a product of natural circumstances, not a product of whether you have a free press. Nonetheless, societies with free elections and functioning press are not experiencing famines, never. The reason is that democratic governments in the sense defined have to face public criticism and they have very strong incentives to avert catastrophes. Now, if you would, it's probably useful to think of Sen's finding, not just as a finding about famines, but a kind of rallying cry about the human importance of output transparency. If you're thinking here of Flint, Michigan, that's a that's a good thought. Okay, so here are some words from President Obama from really the very early days of his presidency. Democracy requires accountability and accountability requires transparency and then there's an invocation of Justice Brandeis's words, sunlight is said to be the best of disinfectants. I hope this will continue to be true. If you go on to OSHA.gov, the government's website, you will see kind of staring you at the face, sunlight about any place in the United States where a worker has died. That's there. Anyone in the world can find it immediately and the company's named and it's happening not quite in real time, but really close to real time and the goal of that was in part to save lives. No private institution wants to be named on the OSHA website as a place where somebody died and no Occupational Safety and Health Administration wants to see a lot of lit names on its own website. So this is about that. The words here are kind of cramped on the board, but the basic notion is that there's a clear presumption in favor of openness, which means that if things are tough, if you are doubtful, disclose it. There's also an idea here which tracks one of the best thoughts in jurisprudence in the last hundred years. It's widely unknown, but it's great. It's from Oxford's Joseph Raz. Raz has an idea called Exclusionary Reasons, which says there are second-order reasons to rule certain first-order reasons off limits entirely. So it's not like we're balancing. We are excluding some reasons for action entirely and there's an exclusionary reason in the president's memorandum. Don't keep information confidential because of embarrassment, because of revelation of errors and failures, or because of speculative or abstract fears. Okay, this has cashed out this idea by saying something that remains a work in progress, which is that the government should take affirmative steps to make information public. They shouldn't wait for specific requests from the private sector. This is super important. The government has a ton of information and if reliance has to be had on a request from someone in this room or an organization or Harvard Law School, that's not good because there's going to be a ton of things that people in the private sector don't know about or don't have the resources or the time or the focus to try to get to the public. Put it out there. Okay, one thing that emerged from the memorandum you just heard about is the Open Government Partnership, which now involves dozens of nations of a very diverse sort and the most noteworthy thing about the Open Government Partnership is the existence of national action plans, which have clear commitments for disclosing things that could uncover corruption or which could show that things in a nation aren't going so great along one or another dimension. Which could promote transparency. So from the US National Action Plan for about a year ago, just going to give you one example, which is the Department of Education is going to use transparency through a college scorecard, which provides a ton of data on costs and student outcomes so that parents and applicants can see it. That is using transparency first as a disinfectance. So places that cost a lot and don't deliver for their students stand revealed as such. That could be an impetus for better performance and also very different from Brandeis' point. Something that will provide people information that they can use in their lives. The CDC, last year and before, has been kind of a champion of output transparency. And you see a few examples just here. Zika can be translated sexually between men pretty much as soon as the CDC learned that, it told people that. So they know it. There's no change in autism prevalence. That's important to know. We now have a lot of information, both about where obesity is on the rise and where it isn't so much, and about how people can protect themselves against food poisoning. All of those outputs are in the public domain. Here's something near and dear to my heart, though it probably won't get the emotions going in quite the same way as the CDC action, which is the Department of Transportation in recent years has revised its value of a statistical life to the point where, as of 2015, it's nine point four million dollars. Now the notion of a monetary value for statistical life, that's not something that people rise up and cheer, hooray. But it's used by executive orders under presidents of both parties, and the American people ought to know about it. Both the fact of it and the number. Here it is. If the number is wrong or if doing it all is contestable, well, everyone in the world gets to see it and to make that argument. For climate change and probably the most important, how many of you have heard of the social cost of carbon? No, one. Okay. Well, here you're going to learn something very significant, actually. When the United States makes decisions about energy efficiency, about regulation of power plants, about regulation of cars, it decides on the level of stringency. It does that in part by reference to the cost and benefits. How do you calculate the benefits of a specified reduction in greenhouse gas emissions? What's the number? The number is the social cost of carbon, and it's centrally important to decide whether you reduce greenhouse gas emissions a lot or a little. If it's two dollars, the level of stringency would be really low. If it's 70, it will be pretty high. The social cost of carbon figures are in the public domain. You're seeing them right here. There's full output transparency and there is public comment thereon. Okay, in the same general vein that is regulatory choices that actually affect people, there is a regulation involving silica, one of the most important occupational safety and health regulations in the last 40 years, probably. Probably the most important in the last 15. Why is the government regulating silica? And why is it regulating it to one extent rather than less or more? Full transparency of the output, which is you're going to get 600 lives saved per year and 900 silica-related illnesses prevented each year. That's just words, but I confess I was in a room with family members for people who had come down with silica-related illnesses or who had died from silica-related illnesses. If we had kind of a few hundred of them marching outside this room, then the human reality would be visible. In any case, the benefits of the silica rule crush the costs and everybody can see that. You can see that because these outputs are in the public domain. Okay, a generalization of much of what you've just seen is the website data.gov. That's the outgrowth of the president's memorandum and basically this is output transparency with a vengeance. The idea is high-value data sets are being made available without a public request. So what you see on data.gov, I believe, is somewhere between overwhelming majority and kind of all of them just being disclosed because the government thinks it's a good idea. The high-value data sets are defined in terms that are kind of helpful for understanding what output transparency is good for. And most of this is self-evident, improving public knowledge of the agency, increasing accountability and responsiveness. But what I want to underline is furthering the core mission of the agency. So if you have an agency like the Occupational Safety and Health Administration that requires people to report in cases in which workers are getting sick so that everyone can see that, then the Occupational Safety and Health Administration's mission, which is reducing disease and accidents, will be served. And there's a great deal of work suggesting that that happens. Transparency does that. Okay, in terms of the data.gov website, here's just a few examples. We have about 200,000 data sets when last checked. It includes information about crime, about chronic disease, about food access. We also have, as a result of data.gov, a lot of private creativity, which is probably on many of the cell phones in this room, you have the consequence, which can allow such simple things as making better decisions about what flights, hotels, and rental cars to choose from and storing your personal health information so you can download it. Okay, here are three empirical notes. Toxic releases in the United States in the 1980s were a quite serious problem. It's not a problem today. 1980s is a pretty serious problem. Congress responded by some statute that had a bookkeeping measure where people had to report the toxic releases to the EPA. It was not thought to be a regulatory tool. So EPA knew the consequence of the toxic release inventory has been to produce plummeting of toxic releases because no company wants to be one of the dirty dozen in California. And the result, unanticipated success story here, has been to reduce emissions. Calorie labels are a form of output transparency where we have information, an outcome of a process, about what the caloric content of food is. The recent data suggests that obese adults in the United States have been losing weight because of the calorie labels. It's working. On your credit card bills, you see transparency about outputs, that is outputs of private and public processes, disclosing the result of filing a paying late or going over. American consumers as a result of the credit card law are saving over $10 billion annually. And the concentration of winners from the credit card law are people with bad credit ratings. Now as inspiring outcomes of law go, I say this rates high because the people with poor credit ratings are the people who are having real economic trouble and they're saving money. Now the credit card law that's producing those savings has a lot of provisions in it, some of which aren't transparency provisions, but the transparency provisions are contributing. Okay, here's a great website, Regulations.gov. It has the beautiful words, let your voice be heard. And I am here to testify to you that my biggest surprise in the federal government. I got 70 surprises, but the one that tops the list is what you're looking at here, which is among the fancy administrative law people, including authors of the leading textbooks, the public comment process is thought to be a form of Kabuki theater. It's not real. It's a TV show. It's on Netflix. It doesn't affect government. That is the sophisticated view I just described. It turns out to be false. The comments that come across on Regulations.gov, they are read, I don't think, by the President of the United States, but people who talk to him every day, they read some of those comments and people who talk to the people who talk to him every day, they really read those comments. So the outcome of what government does is often kind of decisively influenced by what comes across the transom on what you're seeing. This is really transparency working as a corrective to inaccurate decision-making. So our two kind of motivations here are providing people with information that they can use in their lives and increasing the accuracy of government judgments as well as the accountability of government officials. Friedrich Hayek, big celebrator of the price system, which he thought as a marvel because it could aggregate the information held by millions of people. Hayek was not a fan of government action that didn't involve just reliance on markets, but there is something, I think, importantly Hayekian about a process of transparency by which government's outputs are subject to public scrutiny and review. Okay, so in terms of the lack of output transparency being a problem, serious and imminent health risks people don't learn about so they can't do something to reduce those risks, useful information about the costs of care at hospitals are not provided to people, information that the government has about its own failures and imperfections, how policies have gone wrong, how long it takes to get something done. People don't have that, which suggests it's useful to distinguish between consumers on the one hand, choosers of some bit of conduct, what safety seat am I going to buy for my kid, where am I going to live, am I going to go to the airport now or 20 minutes from now, or people acting as citizens or watchdogs and in either case we have problems with the lack of output transparency. Okay, this is a kind of fussy slide for which my apologies, but the idea is that the amount of government the information has that counts as an output is not infinite, but kind of in that direction. So you should probably be thinking about outputs that are useful to people in some way. If there are outputs whose disclosure would risk national security or personal privacy, then there's a reason not to disclose them. And if it's an output that is of no consequence for anyone, then it's not that important that it come across, but we remember the presumption. Okay, that's the kind of every bone in the body enthusiasm part of the remarks. And the hope is that we will see a ton of output transparency in the next administration. And if there isn't, there's going to be some explaining to do. Okay, input transparency, I'm going to try to suggest as a more mixed bag. So you're seeing the situation room. This looks like a discussion of foreign policy. A lot of people are saying a lot of different things undoubtedly. If the room is working well, people aren't in agreement. And now you're seeing the Oval Office where there's a discussion of domestic policy, maybe the budget. And you're seeing a smaller crew, but if the process is working well, those four people are not in complete accord, or they are at least referring respectfully to the views of people who aren't in accord with them. Okay, so the budget deliberation process, I don't know for sure, but that could have been a budget discussion there, maybe. The budget deliberation is not transparent. So if someone from the Domestic Policy Council says, I think that the EPA's budget should be increased by 150 million, and someone from the National Economic Council says, I think it shouldn't be increased at all. Someone from the Council of Economic Advisers says, I think it should be cut. The American people don't learn that. That's not public. Regulatory deliberations, it's a more complicated story. There's more transparency, but as things fall out, it's not so far apart. Okay, so Senator Warren, Harvard's own and someone I greatly admire, is concerned, and I think quite reasonably, about EPA records showing that industry groups have a monopoly over informal communications. Now we're talking here about inputs. It's not intra-government inputs. It's inputs between government and the private sector. She's exercised by the fact that the industry groups are talking a lot more with EPA than the public interest players. Now, that apparently fact is itself public, but it isn't always who an agency is talking to. Okay, that's a discussion. It's not a very sensitive discussion as it happens. It was a discussion about regulation and its costs and what things are looking like in a tough economy, but if I told you more, I'd be kind of ashamed of myself, even though it wasn't particularly sensitive, and that's connected with what I'm about to say. Okay, so my thesis is that typically the benefits of input disclosure are really low. So if the head of the Domestic Policy Council disagrees with the head of the National Economic Council, who disagrees in turn with the head of the Council of Economic Advisers, it's America benefit from hearing that. People in Flint, Michigan get something good out of that. There was one time in government where the White House gave a lot of transparency to the New York Times about inputs, and they told me, sit and talk to the New York Times. I'm going to name the guy John Broder. Does anyone here know John Broder? Okay, so my little boy, who's six, had a line for a while. He would look at someone and he would say, with great mischief in his eye, he's 67 now, he would say, is he a bad guy? Say it with delight and glee, infectious delight and glee. I think about John Broder, and I think, is he a bad guy? He's a great reporter, and so far as I know, he's a great guy, and he was given complete input transparency, and he mangled the story really badly. He described it as a tale of capitulation to the Republicans in order for President Obama to be reelected. I was in that room on multiple occasions. There was no such thing at issue, even a little bit. It was completely a merit's decision, and it got everyone focused on what was not true, and if it was true, there would have been some importance to it, but the really important thing was the decision right. It was a substantive decision, and even if it had been a politically motivated substantive decision, this was the substantive decision good. There's a very good reporter for the Washington Post, and I'm going to be a little uncharitable to her right now, so take this as kind of a version of her that's less good than the real her. She's a tremendous reporter, and she kept asking me after I left government, who said what to whom? Who said what to whom? She wanted to make someone look like a demon, and someone like a good person, or the palace intrigue, and I thought this person really cares about the environment. Write something about the environment and policy. Write something about an output. Don't try to figure out did Larry Summers and Carol Browner disagree on something. Does that matter? For historians, does it matter a little bit, certainly if it involves the president? For most of this stuff, no historical interest. Robert Gates, Secretary of Defense, wrote a book. It caused a lot of news because he trashed his former bosses. He disclosed a lot of inputs. Most of that, no historical interest. I'll get to that. I hope we'll have time for questions. In terms of the costs, the thesis is the costs of input disclosure are not really low. It deters candor for one thing. The fact that people can't talk by email is a cost. They can talk on the telephone, and they do. A speculation which I think a lot more work needs to be done on is that the risk of input disclosure has much worse consequences than deterring candor. The Office of Legal Counsel has become, decreasingly, a relied upon entity, even though the Office of Legal Counsel is the repository of terrific legal judgment within the executive branch. One reason I think is, if OLC gets engaged, you're going to have some public accounting of what the Justice Department thinks, and that's troubling. There's going to be, from the executive branch, that's troubling. It may be that there's going to be disclosure of stuff, or that the executive branch is going to be trapped by what one lawyer's thought on one day, and that is not something that's welcome. And so the deterrent effect of transparency is that OLC is not talking. Not asked to talk. I get the feeling you're not loving these remarks. I confess that this part of these remarks, I think you like the other parts better. I confess that six or seven years ago I wouldn't like this part as much as I now do, but the adverse effects of input transparency are high. And typically what you get out of input transparency is gossip value, and the Washington Post version of the National Enquirer, and not anything that's going to save a single life or make one American enjoy economic opportunity or escape poverty. Okay, there is a diversion for the public in this stuff frequently. Should the calendars of cabinet heads be public? Would that be really great? Is that what America most needs now? Who it is that Secretary Kerry is meeting with? That's not very important. Okay, the more sunlight that shines on agency processes, says Senator Warren, the more likely it is that an agency's final product will reflect the public interest. A good start would be disclosure of all the meetings between the agencies and interested parties, both before and during rulemaking. That's an interesting suggestion. It's not clear that it's wrong, but one kind of cautionary note is frequently if an agency is meeting with an industry a lot, it's not doing it because it's pushed around. It's doing it because it's going to do something very adverse to the industry's interests, and it wants the industry to understand exactly why. So the point of the zillions of meetings is not to say industry, come please come capture us. It's to say we're going after you hard, and we want you to know why we're doing that, and be prepared. This is what it's all going to look like. And we're listening to you, but the listening most of the time produces no movement on the government's part. Sometimes the government learns something. It instead gives people a right to be heard, which they deserve. Okay, just about done. Let's tidy up a little bit. Many FOIA issues don't fall clearly within the categories of inputs or outputs. Sometimes they're at the border like a report on carcinogens that might be the basis of regulation. So that must be acknowledged. What have we learned since 2009? Here are just four points. Output transparency is hugely important, and we need a ton more of it, accompanied by the presumption in favor of openness. The Open Government Partnership is massively promising. If anyone here is prepared to do some work on it, bless you. It should be a presidential priority and an international priority. I've looked, I can't find much sustained research on it. It would be fantastic to do. Input transparency may be a bad idea, and it isn't necessarily a great idea. We spend in the transparency world, I think, far too much attention on this relatively low-level issue. We should understand the Freedom of Information Act through a welfare lens, which would lead us to be careful with abstractions about legitimacy. Frequently, we're not going to have the data to think about the benefits of transparency and the costs, but at least if we ask about the human consequences, which are really high in many of the cases discussed, think about the toxic release inventory, what's on the OSHA website, calorie labels, which are saving not only pounds but lives. But at least if we focus on the questions of what are human beings getting, we're asking the right question. And a lot of the time, that's going to get us to the right answers. Thanks. Hi. I'm going to propose another potential benefit to input transparency. So with the incoming administration, there may be some worries about potential criminality within the executive branch. And I wonder if you think that if folks are more worried about this stuff getting out, it may give them more of a backbone. If they're worried about history, maybe then that would give more of a push as opposed to these arguments sort of disappearing into the ether. It's a fair point. So bracketing the incoming administration and it actually cuts against my argument in a really interesting way. I think my argument depends on the assumption that criminality will be very rare and the democratic processes are a sufficient constraint. If you lived in a regime like Putin's Russia, let's say, where kleptocracy is present, there'd be an argument for input transparency. So it's an interesting point because maybe the division between the input transparency people, people like it, and people who like me are suspicious of it, is that the former fear, illicit acts, I don't, under Bush and Obama, I didn't see anything biased witness here on Obama. But I didn't see any decisions which reasonable people could reject. Certainly Republicans, nothing illicit. And my understanding is under President Bush is exactly the same. You don't like the decision, but nothing corrupt. And if you think that that's just an artifact of the last 16 years, then the argument for input transparency strengthens. It's a fair point. Let's take another angle on input transparency. I'm here. So in the earlier panels, we talked a little bit about the TPP and the failure of the TPP. And there's at least a good argument to me that you can lay the blame for that failure on the lack of input transparency, a public that doesn't feel involved with those deliberations, lacks confidence in the output, to the extent that we don't even get to an output that can really be made transparent. You have thoughts on that? Well, I'm not an expert on this, but, which makes me about to say hazardous. If the failure of TPP is a result of a lack of input transparency, I will eat this computer. Now, we need a good oven for me to, I just don't believe that. The arguments against TPP made by the left and the right, aren't, they didn't tell us enough. And if they, if there'd been more transparency about it, probably meant it worse, because the transparency would have made people seize on the things that they were suspicious of and say, oh, you talked about this and then you did that. So input transparency, often in a highly polarized environment, creates more agitation rather than less. Now, you, I will also eat this computer if I know as much about TPP as you do. So my guess is, you know a ton more. So I might be wrong on this, but given the politics of TPP, I think there might be other examples, but probably that the transparency that relaxes people is when you're given account of why the thing is being done. So if you say, you know, we're going forward with the silica rule, which a lot of people hated, because it's going to save 642 Americans, but other is die. And the science is solid. And the cost is going to be 900 million, let's say, and by the numbers everyone uses, 600 lives worth 900 million. That should come, but then where are people going to say that? If you say the silica rule, you know, the labor union said this and the Chamber of Commerce said that, right? Or the Council of Economic Advisers was a little nervous and the Department of Labor wasn't nervous at all and they won. To my knowledge, that's not true, by the way, but that would aggravate the where the Department of Labor wanted to go more aggressive than it ultimately did. This is just, this makes people not feel confident, makes them feel, it's a fog, a fog of diverse opinions, where if you tell them what was done and why, maybe? I wonder if you could elaborate a little bit more on the Flint example right here, because I'm thinking that the deliberative process of the who knew what, when, when they made the decisions to essentially use the water systems that they did are incredibly important for the public to know. And setting aside the criminal element here, there should be some political accountability there that the public has a right to know. I'm thinking of a stylized version of the Flint problem, which may not be actually the Flint problem. So the thought is, if the government knows water's unsafe to drink, it should tell people as soon as it knows that. If there's, it gets harder if there's internal disagreement about the safety issue with excellent people saying, you know what, there isn't a safety risk, then you're in the deliberative process. I'm using it as a case where the government actually knew who in the government I'm bracketing, but the government official actually knew there was a health risk and didn't tell people. That's a lack of output transparency. So if there's knowledge that people in some part of the country are in danger from, you know, terrible air pollution that month, then to hide that or not to disclose that, that's kind of the core of what government can't be quiet about. And I'm understanding the Flint case in that way. I guess if the Flint case, it's a hard one for me, if as you're suggesting, what government knew was very unclear, lots of people had different views. And I'm not sure what to think about the category of cases where some people think people are at risk and other people disagree. It's not clear that there's an obligation to disclose there because it's influx and the people who think there's a risk might just be scaring people. If you have a political appointee who is overriding the decision of a scientist, you know, that disagreement is important to note, I would say. And I think that would fall squarely in the input transparency. Okay, you raise a very interesting, that's really interesting. So we can think of cases where there's a kind of science policy issue where it's not clear who gets to win the scientist or the policy person. But if there's a pure question of science, sounds like if the scientific judgment uncontradicted is there's real risk here, then the government is obliged to disclose it. And whether we call it an input or output isn't obvious. But I agree with you analytically, you should put it on the output side of the line. Two things on TPP. Thank you for your question. I don't see it as a failure, I see it as defeat. I think it was defeated. And I disagree with you, I think a lot of it really was. I don't think a lot of people really were able to read much of the details. I remember going to a meeting up in Maine early on in the process and legislators in Maine were asking for details of the agreement as it was just beginning to emerge and the issue of not being able to get access to the content and the argument that it was corporate representatives who dominated to the detriment of the public interest I think was huge in the view of the direction of TPP. So the openness was I think a crucial element. And the second thing is I guess I disagree with you also about when you talked about Elizabeth Warren's statement why shouldn't people know if there's no harm if it's an innocent meeting to tell people you know hey we're about to clobber you why shouldn't the public know that and how are the public going to know whether you're talking to people because you're going to clobber them or because you're not going to clobber them. Okay so good thank you for both. On the first I'm not ready the computer yet but but I hear you. On the second I didn't mean to disagree with Senator Warren on that so the office which where I worked the practice is always to disclose the fact of meetings. So you know it's it's not a bad thing but and it might even be a good thing but okay so the Office of Information Regulatory Affairs and the Executive Office of the President discloses all the meanings with with people. Now has one child in the American in our beautiful country of 330 million people had a better life because of those disclosures? I don't know either. Well I just okay so I'm for it so I'm for it but you know I guess another way to put it is how to get at this. 480,000 Americans die from smoking every year 28,000 die from on the on the highways that's real. Okay now if we could have transparency policies that would make a dent in those things that would be something to get really excited about and for the transparency community that to focus on that is very worthwhile. The Oira's you know I'm being very candid here Oira's multiple countless disclosures of meetings I'm for it but you see the point is there one person in Utah who's had a better life because of those disclosures? I don't think so. So I agree I think I agree with you and Senator Warren but it's but I like much better the the idea before that if the scientists have consensus that there's a risk let's characterize that as an output and say that the government should be disclosing it even if no citizen is asking for it. I think we'll have to leave it there. Our next panel is going to come up in a couple minutes. Thank you so much Professor Sensen. Thank you all. Thank you. Thanks to all of you.