– why open data needs a non-commercial-use license, and the lessons of microcredit.
“Government 2.0” is the initiative to make government data open to the public using Web-based technologies. Leading light Tim O’Reilly describes it as “government as platform”. The idea is that open data – provided in such a way that programmers can write software to read it, analyze it, and transform it – increases transparency and promotes innovation.
Government 2.0 is on a roll. It got a big boost in the US from Barack Obama’s early memo on transparency and open government, and the setting up of the data.gov website. In the UK there is data.gov.uk and David Cameron’s “Big Society” initiative. Even Canada’s notoriously secretive government is consulting about it and many cities have opened up data feeds. And there are other initiatives around the world. Sounds great? Well yes and no.
The rhetoric of Government 2.0 draws heavily from efforts by private citizens and non-profit groups to make government more accountable. It has a civil liberties flavour, with talk of citizen engagement and of citizens’ rights (“giving citizens access to data that is theirs”), participatory society, collaborative democracy, transparency, and so on. Take the new collection of essays Open Government. The examples of Government 2.0 initiatives almost all deal with citizen access to the US government’s inner workings: campaign contributions, lobbying data, congressional votes, legislative proceedings, federal government contracts and spending, court proceedings. It describes efforts by non-commercial groups such as opensecrets.org, maplight.org, followthemoney.org, govtrack.us and so on to use this data to enforce greater accountability. This is all fine and good – although proponents should recognize that this access to information is one step in an arms race and that those who want to hide information will now look for ways to do so.
But the “Open Government” project has a second agenda: it demands that data be made open not only to citizens, but also to private companies. The recent Gov 2.0 Summit and Gov 2.0 Expo, also organized by O’Reilly Media, are big Washington events sponsored by major technology companies. It is clear that there is money to be made in Government 2.0 – money that is mentioned very sparingly in the Open Government book.
Participation or privatization?
Should we care if private companies get our data for free? After all, maps and weather have already been made available, and innovative commercial applications such as Google Maps have made it useful to the public. Many Government 2.0 enthusiasts see commercial opportunities complementing non-commercial activities. Tim O’Reilly writes, “The whole point of government as a platform is to encourage the private sector to build applications that government didn’t consider or doesn’t have the resources to create. Open data is a powerful way to enable the private sector to do just that.”
There’s another name for outsourcing government services to private industry, and that’s privatization. Talking of “a dramatic redistribution of power from elites in Whitehall to the man and woman on the street”, as David Cameron did, is a lot like the Margaret Thatcher line that selling off nationalized industries was restoring them to the people. As Ed Miliband says, “for all the talk of a big society, what is actually on the way is cuts and the abandonment of community projects across Britain.” O’Reilly is very close to supporting the Big Society program (he praises it in a recent tweet) and says that “In some sense, government hitting the wall on deficits is a good thing, if it forces a real reboot,“ which is not only callous, but uncomfortably close to Naomi Klein’s “Shock Doctrine” thesis. You have to wonder, is this what the civil libertarians who have been pulled into the Gov 2.0 effort really want? Some of them need to think a little harder before following Tim O’Reilly’s tune.
The problems with profit
Many of the heavy hitters in Silicon Valley believe in “social entrepreneurship”, in the belief that they represent an enlightened, humane, and smart capitalism. But money has a way of corrupting. The story of microcredit has lessons that Government 2.0 promoters could learn from.
Mohammed Yunus won a Nobel Prize for founding the non-profit Grameen Bank that extended “microcredit” loans to groups of poor people, with very encouraging results. When Pierre Omidyar, the eBay billionaire, got involved in microfinance he decided that the “non-profit” part of the story was a limitation. As the New Yorker’s Connie Bruck reported a few years ago: “Yunus is now seen by Omidyar and many others as the archetypal founder, too wedded to his original vision. In recent years, younger and nimbler players have been taking microfinance—their preferred term—toward the idea of building a fully commercial, profit-making sector. This conflict, between pure do-gooders and profit-minded do-gooders, has come to define the current debate in the microfinance world.”
Unfortunately it has not turned out so well. Yunus’s worry that profit-oriented companies were “pushing microfinance in the loansharking direction” has been born out. This year’s IPO by SKS Microfinance, a for-profit company active in India, made tens of millions for some of its board members, as well as for board members of Seattle-based nonprofit Unitus, which has invested in SKS. Meanwhile, the loans from these commercial microfinanciers have become “death traps”, and after seventeen SKS clients in the Indian state of Andhra Pradesh committed suicide the clamour for answers about the role of profit has become louder. There are real differences between profit-seeking and not-for-profit groups, and they must be kept in mind.
A step forward
There are other issues with Government 2.0 that I may come back to later, but let’s wrap this up. Glossing over the difference between companies and citizens obscures key issues at the heart of Government 2.0 and risks corrupting the whole enterprise. Fortunately, there is a solution. When government puts its data into the open, it does so with a license attached to it. In most cases this license permits both non-commercial and commercial use of the data, so long as the source is acknowledged. It’s time to consider licensing the data for non-commercial use only, and having a separate license for commercial users. Perhaps permitting some limited commercial use for free, but charging for more extensive use, would be a first step to protecting our data from the temptations of profit.
Update: Part two is here.
Nice post, Tom. You might like to check out this post, specifically the discussion of the case from India: https://gurstein.wordpress.com/2010/09/02/open-data-empowering-the-empowered-or-effective-data-use-for-everyone/
As to the licensing idea, I’d like to see this played out more extensively. I’m so sure what you briefly suggest is nuanced enough, but instead perhaps the distinction shouldn’t be commercial vs. non-commercial but some stipulations (such as privacy, data portability, etc.) that could seek to limit the ills of private power without blocking all profit.
Thanks Kevin. Michael Gurstein’s post is a good one; a similar point is made by Kentaro Toyama in his recent talk on myths of ITC4Development http://www.youtube.com/watch?v=E_mTwm5m8DM (see Myth 1 at 8 minutes).
I had originally thought I’d go through four problems, of which the others would be privacy, the fact that internet access is not democratizing (your point), and ignoring the fact that information is often not the bottleneck in improving government services. But then it got late, and the post is long enough.
I’ve been enjoying your blog – a balanced and open-minded point of view on technology matters makes a nice contrast to many commentators (me included, too often).
Oh yes, and I agree that the licensing question needs more thought than I could give it. I’m sure there are quite a few options, and putting limits on specific uses may well be a good way to go.
I think your criticism of microfinance and, generally, government “2.0” is correct.
But regarding commercial/non-commercial use, I think this is a mistake. In fact, this is what a lot of commercial enterprises do when releasing software, and at times data. It is made free for non-commercial/educational use, so that others may not compete.
Take free/open-source software, which we have a lot of experience with. The Free Software Foundation’s definition of free, and the Open Source Initiative’s definition of open source both require a license NOT to preclude use, e.g. for commercial purposes. Non-commercial use is vague and very restrictive. And it forbids use that perhaps it shouldn’t, e.g. use by individuals for a commercial purpose, use by a small collective enterprise for a commercial purpose, etc.
What is, I think, useful, is the concept that the free software community calls “copyleft” and Creative Commons calls “share alike”. What this concept means is that if you enhance the software, or re-use the work in some way, you must release your changes under the same license. This keeps the work freely available and means that, even if the changes made were for commercial purposes, others can use those changes for whatever purpose they like.
Finally, in the U.S. at least the point is probably moot; generally government data, publications, etc. that are created by the federal government are public domain, and this is not going to change.
Large corporations with the latest’s technology used to have a have the competitive advantage. In today’s economy, having the latest technology doesn’t only give us the advantage, it levels the playing field for IT is a necessity for any business to compete in today’s local, yet global economy.
For Micro-Finance survival, they need to muzzle their Spin Doctors and listen more to their High Priest
A string of suicides in Andhra Pradesh that put micro-finance under the spotlight, triggered a backlash because of which, MFIs found themselves reduced to fighting for their basic survival. No surprise here to find a variety of spin-doctors functioning as their apologists, fending off and neutralising any criticism that the industry faces currently, almost oblivion to the fact their support is to a slow sinking Titanic. Two of the most significant spins in this debate are those related to suicides and interest rate. In this post, we bust these spins.
“I believe in Schumpeterian creative destruction. Its time has come. The present MFI model has to go…. It wasn’t just about giving loans. It was also about creating livelihood mechanisms, which would build capacity among the poor to repay their loans easily, and leave them better off than before”
This is Economic Times quoting Vijay Mahajan, considered the high priest of Indian microfinance suggesting that either MFIs change their business models or go bust.
Read more: http://devconsultgroup.blogspot.com/2010/11/for-micro-finance-survival-they-need-to.html
Good points. I’m not sure the analogy between open source code and data goes so far. As I’m sure you know, it is quite common for photos and other creative works to be placed under a non-commercial license. I do agree the word “non-commercial” is vague though – at least, I haven’t seen Creative Commons define it clearly.
The open data initiatives are going beyond publications, and there have been issues with things like lists of employee names and dates of birth (Seattle had to release these to a TV station), which have given rise to concern. Are these public domain? Not sure. Maybe limiting re-use in some way would help address these isses.
Thanks Rajan. Interesting posts.
Sharealike is a good start however can be circumvented via advertising inclusion. NC in Creative Commons is a fairly balanced approach to Non Commercial use and may cover the bases better.
Also note that Gov 2.0 does not mean open data alone, it includes the application of Web 2.0 tools and techniques to public governance practices, which also has wide application.
Hey I like your idea about the open license, Tom. Like Erik, I don’t think the “non-commercial” part is helpful but a copyleft GPL (Affero to get around the network service loophole) license would be a great fit for public data. I’ve never thought about Open Source licenses applying to data. It makes a great deal of sense, and I’m not a fan of Stallman and copyleft in general.
It makes sense for commercial data sets too. OpenData 2.0 🙂
Under US copyright “facts” are public domain. There is a well-known case having to do with the phone book, which was judged to be public domain. What the difference is between this &, e.g., maps is not entirely clear to me.
I didn’t mean to imply that there is necessary an analog between software & data. Each of these communities will need to work out what “open” or “free” means to them. For instance, while I think it would be ridiculous to release software under a “no-derivatives” license, releasing a novel under a no-derivatives license would make sense to me.
I do think a critique of these initiatives would be useful. It is absurd to think that releasing government data is going to make a huge difference in the way governments act. Certainly reliable information about government crimes, malfeasance, etc. is important, it may even be necessary, but it is not sufficient to make significant changes.
Yes, there are a lot of commercial data sets that I would love to see. It would be great to get a look at real sales figures for Amazon or iTunes, but pigs will fly before that happens.
Thanks for clarifying about Gov 2.0. You don’t have a reference to Creative Commons’ definition of NC do you? I have not been able to find one.
And of course reliable information about government crimes is not going to be released anyway, not matter what open data directives are handed out.
Do you consider this a definition? It seems vague to me, but I guess it works.
You may not exercise any of the rights granted to You in Section 3 above in any manner that is primarily intended for or directed toward commercial advantage or private monetary compensation.
It seems that CC did a report on how people interpret that phrase you quote last year: http://creativecommons.org/weblog/entry/17127.
One commentator summarizes the report this way:
Another factor to consider is that “open access” to public information does not necessarily mean it is equally open to all social groups as a practical matter, and in fact acts to the detriment of the poor:
I agree completely.
There is a lot of attention given to “Open311” services (http://open311.org/), which let citizens report non-emergency issues to local government. The archetypal case is potholes that need fixing. Personally I doubt that information is the bottleneck here. There is no group of city workers sitting round waiting for a call about a pothole so they can go out and fix it.
If these services come into widespread use, my guess is that they will skew services towards neighbourhoods where people report most frequently, which will be the smartphone-owning home-owning better-off neighbourhoods.
Thanks, Tom. I wish I had more time to write, but I’m glad you’ve started doing so more frequently since I learn a lot from each post.
I’m not sure I understand your argument exactly. Is it basically:
If that’s the argument, I have two problems with this:
Incidentally, nothing against anyone who actually enjoys getting effed in the ay.
Sunny: obviously you are right that privatization of government services is nothing new. What bothers me here is that the campaign for privatization (to oversimplify) is cloaked in a rhetoric of radical openness and power-to-the-people populism. I don’t like lefty language being used for what may be right-wing ends.
I do like the point that when open data services get created, competition may help avoid the public getting effed in the ay as you so daintily put it. But where does this Magic Card contract thing come from????
I’ve never really understood “left” and “right”. The “right” want the government out of the way, but they want to be in the way when it comes to, say, gay marriage. OTOH, the “left” wants equality for all, but will gladly entrench monopolies and advantage the wealthy. So given that “left” and “right” mean nothing to me, here’s my attempt at a response:
I sincerely do not believe that open data is a “campaign for privatisation”. It may be happening given the circumstances, given game theory, but I don’t think anyone out there is saying “great, now there’s open data, we can use it as an excuse to privatise stuff” — and that’s despite what people are saying regarding private enterprise picking up the slack that the government won’t.
There’s nothing inherently wrong with empowering a business with data. Giving it away for free might enable smaller players to compete with the entrenched businesses. If this data weren’t free, it’s not as if MegaCo wouldn’t buy it. It may even prefer that, because TinyGuys won’t be able to afford the license. You could make a proviso for non-commercial use, but People developing applications also have to eat.
While you don’t agree that Erik’s comment regarding Open Source applies here, there have been various debates about whether to make code (effectively) “non-commercial”, or attribution only. The problem with making code non-commercial only, is that it doesn’t see nearly as much activity as the attribution style licensing. This is because Open Source is often used (quite heavily) by commercial companies, and they’ll need to add a feature here, or fix a bug there. It also evens the playing field: It allows smaller companies to compete with the entrenched players like IBM and the like. I have a feeling open data is in the same boat.
The idea behind Magic is that you have a very simple rule set on top of which there are “exceptions” (which is what’s written on a particular card). So a contract would basically be a set of Magic Card (equivalents) where the government has basically checked the fairness of each individual card. This allows a business to add cards (thus allowing for a number of options for a contract), but the cards are written in a way that a lay-person can understand them (with colors and other such means, for example “green” for a “good clause” and “red” for a “bad clause”) and can build an understanding of the contract through seeing the same sort of “patterns” over and over again.
In fact, I was silly not to refer to the Creative Commons licenses, where the license is spelled out using icons and other plain language, and you can mix-and-match the icons. All “contracts” (liabilities) should be like that. I think this would preclude a business from damaging individuals, thus rendering your whole argument void (well, in this article, not the follow-up, which makes total sense).
I don’t really think that the Open Government people are out there twirling their moustaches and planning to divert all this goodwill to mischievous ends. I do think they don’t see a conflict between profit and social good – or even a tension in some cases – and I do. So it comes down to a political position as much as anything.
And I suspect you are right about Open Source. The GPL doesn’t get rid of commercial use, of course, but it does limit it as a side effect. I wonder if a “GPL-data” license – requiring any data it is combined with to be released under the same license – may have the same kind of viral effect?
(I didn’t know you’d started blogging again, Tom. It’s taken me all evening just to catch up.)
You write: “There are real differences between profit-seeking and not-for-profit groups, and they must be kept in mind.”
Somewhere in my catching up I think I came across a post in which you expand a little on the distinctions between the two but I can’t remember exactly where – maybe I just hallucinated it out of the pattern of weft and woof of your posts. Whatever, if you’ve ever a yen to explore it, it’d be nice to see a distillation of those real differences.
By way of a weak quid pro quo, the commercialisation would appear to be already under way in the form of care home and schools chooser “apps” based on HMG’s open data releases:
There’s a dozen pages of listings of “apps”, the vast majority of which would seem to be not-(intended)-for-profit but there does appear to be the beginnings of a separation between the n-f-p and the p-s in terms of chosen domain (care homes, schools, home-buying) – i.e. the low-hanging fruit is already being commercially harvested and it’s difficult to see a rationale that would prompt a subsequent n-f-p entry into what are already commercially-dominated areas.
Poking around behind the scenes of some the listed commercial “apps” suggests that the big boys haven’t arrived yet (or don’t see enough mileage in it). There has been some speculation (perhaps just perennial optimism) that there may be significant business opportunities in the analysis and presentation of open data, given the increasing emphasis on providing “consumer choice”. There’s certainly room for improvement: I’m picking my way through an analysis of the metadata available for the 4,600 datasets and it’s not pretty.
We’re talking about privatizing services that the government does not currently provide. Surely this distinction is critical in evaluating the value of these new private services?
That said, I think idea of a two tier license with non-commercial and profit-seeking treated separately (with a generous free license for small businesses or new applications) is pretty great. It’s fair, and more sustainable than the alternatives.
Full disclosure: my nonprofit employer, Global Integrity, is funded by the Omidyar Network, mentioned in the post. Small point to consider: most (all?) of the support provided by ON’s governance program has indeed been grants to 501c3 nonprofits or their overseas equivalents.