Breaking down public goods: with an idea about privacy

Many things are provided as public or collective goods that don’t need to be. We don’t need to provide hospitals or schools as public goods. We could provided them on a full choice, fee for service basis. But once we get to providing safety nets, minimum standards or free goods it gets trickier. Free goods don’t come with price disciplines, so at least on the question of the cost of services, choice doesn’t work too well. So where we are providing state based services, you take what you’re given from within some socialised choice set. These days you can choose schools and hospitals to a certain extent but they’re choices between services which cost about the same amount.

Laws tend to be public goods, but it is possible to inject some ‘contestability’ into them – with competitive federalism and charter cities. I’m not saying that these things are preferably to socialised supply, but they can introduce variation and choice in ways that can be beneficial. And they also make good ‘thought experiments’ to help us unpick what is and what is necessary in the public good of laws.  In each case however the public good – within the state, city or country where it applies – is still a collective good, with contestability being introduced by people voting with their feet.

In the heady early days of Demos Charles Leadbeater waxed breezy and lyrical about how we might be able to choose our own social security system – like you choose your level of insurance in a private market.

Anyway I was at a conference on data linkage in education recently and was struck with how degraded our information systems are because of concerns like privacy.  I don’t want to belittle privacy concerns.  But some thoughts I had about it were these:

  • with codes of privacy protection written into national and state legislation being collective goods, there’s little room for tradeoffs. One just has to deliver a high level of privacy and even the substantial risk of de-anonymisation is enough to prevent certain data linkage projects – or to ensure that the resulting resources are unable to be openly released and so subjected to analysis from unusual and potentially rewarding perspectives.
  • There’s no price discovery.  Given that privacy is a valuable private good, the real question is ‘how valuable?’. Here it’s not easy to know because people might say one thing in a survey and reveal different preferences when called upon to choose. This isn’t me being theoretical.  People are generally not all that happy about firms knowing all about their shopping habits. (though I’ve quickly looked for survey evidence of this, but haven’t found it.) But the real question is ‘how unhappy?’ And the answer is, a few cents in every $10 whic his all you have to offer most people to happily allow their data to be tracked in return for some baubles very year or so from accumulated frequent buyer points.
  • So where the data is commercially valuable we have a bit of a market going in privacy and non-privacy. In the public sector, not so much.

So I wondered aloud to the group whether it might be possible to allow people to opt into a looser set of privacy protections in return for some low cost reward. As I said to them, I’m really not too concerned if my results from Harkaway State School are available for people to analyse. it just seems crazy that so much insight can be gathered from linked data sets – and yet it’s so difficult to do it because one needs to adhere to something close to the privacy gold standard throughout.

The other thing one could do is to focus privacy regulation more closely on the publication of private details, rather than the potential discovery of them. Thus one would be more permissive of privacy issues for public good research but one would still make it an offence to publish such information in a manner that identified people. We don’t stop court cases taking place, and sometimes judges leave courts open but then suppress any public disclosure that would enable identities to be discovered publicly.

Is this practical?  Probably not, certainly in the immediate future as far as the politics is concerned.  And the proposal to seek suppression of the publication of private details would not stop rogue sites run from other countries from circulating the information – though it’s hard to see an interest in doing on a systematic scale.

Anyway, I mention it as a thought experiment.  Somehow it’s a great pity that everyone gets the gold standard of privacy protection as the default, rather than be given privacy protection when they really really want it – or are prepared to tick a box saying they want it.  Is that so much to ask?


This entry was posted in Economics and public policy. Bookmark the permalink.
Notify of

Newest Most Voted
Inline Feedbacks
View all comments
12 years ago

I agree with you, in a free country why can’t you sign up for an annual full cavity search. But there are various ways in which we are not a free country, and one of the freedoms that the govt usually wants to constrain is the freedom to be stupid about certain things; like buying from investment promoters and telephone sales people.

I guess a big part of the rational for privacy laws is the fear that people will be stupid with their privacy and give those nasty companies our shopping habits, or our phone numbers. Once you’re being protected from yourself, well, how can they let you opt out?

12 years ago

I agree. Shouldn’t you be able to devise some kind of frequent flyer points for public data usage?

I.e. every time someone accesses the database(s) in which your particular data is stored you earn x frequent flyer points…?

Or in the specific context of schools you could link it to a school ‘vouchers’ program with ‘voucher expenditure’ being able to be tied to a FFs type program.

Or Medicare cards could have FF features built in. People spend a lot on medical services and happily tell their credit card proi

I think you would get very quick support for your conclusion that average person’s value of privacy is quite low.