Public records and privacy

Often, these days, I find myself looking at web sites about politics, finance, and government and saying, "wow! that's cool... but what is its effect?" is one such site. It's worth a look, because with it you can see what your neighborhood's politics look like. However, don't be surprised when you notice that your name, address, occupation, party affiliation, and amount of contribution are sitting right there on your screen.

This is another of the growing number of public records mining systems that are providing unprecedented access to public information, but they have also changed the assumptions that we have for what level of privacy is available.

One of my favorite examples is the Fairfax County Department of Tax Administration's Assessment Search site. This allows you to look at every piece of property in Fairfax County and find the last sale price, current assessment value, number of bedrooms, dimensions, etc.

Of course, all of this information has been available for decades if you were willing to get off of your duff and go down to the county offices during office hours, wait around for a clerk, and then dig through the volumes of data to find the information that you seek.

However, now, in many jurisdictions, this information has been put online. In the interest of allowing "customers" to provide self service, all of this information is now available free for the taking by just typing a simple address or name.

The continuing question with all of this data availability is whether we are getting to the point where the amount of data available in "public" records is really appropriate. Two classes of people seem to think that it is not: privacy advocates and data providers. Clearly, they have two different ideas of how and when this data should be kept secret.

Privacy advocates argue that the data should be available only in aggregate, i.e. names and most identifying marks should be unavailable. The idea, similar to those in some countries in Europe, is that all information about you should be owned and controlled by you. It is appropriate, they believe, to require consent for the release of any information.

Data providers believe that information shouldn't be available in aggregate or in bulk, because they used to provide that information to other people on a contract basis in the past. Clearly, their issue isn't with the confidentiality, but with the idea that the government would give it away for free, when there is money they could make selling it.

You won't be surprised to hear that I'm not in agreement with the data providers. From my perspective, if you're going to make public information available to anyone, it should be available to everyone.

But, the position of the privacy advocates is more interesting. At first glance, the idea that information about you being only available when you choose to disclose it sounds like a great idea. The problem quickly becomes apparent when you consider somebody other than yourself.

If the "you" is a convicted sex offender moving in to your neighborhood, most people don't like the idea of the felon being able to determine if his/her neighbors know his/her past. But, that's an extreme example.

How about the example of the Fairfax County tax records? It seems dangerous that you can type in my address and find out that I paid less for my house in 1993 than the previous owners did in 1989 (true story, remember that property values can go down, even here). Certainly, that isn't anyone else's business. However, if I were buying a house, I would like to know how much people paid for the homes in the areas that I am searching in. Simple solution: just rub out the owner names, right? It doesn't secure my information from anyone who looks me up on Google or in the phone book, but at least you aren't given my name when you look up the street address.

But, even that has some interesting problems. Consider, if you will, that you are living in Manassas and a man shows up at your door, completely unannounced, offering to buy your farm for your tax assessment value + 10%. Sounds like a good deal, but you're curious why this sudden interest in your small plot of land. For some of the people who were living in that area in the mid 1990's, a few interested people were able to figure out that the land was being bought systematically. It turns out, the Disney Company was acquiring land for use as a theme park. It never got built, due to lack of popular support, but without access to recent sale information, the Disney Company would have been able to buy all of the necessary land without any indication it was going on. (As it stands, they did most of it under cover of third parties, and the only way people noticed it was that a large percentage of the area was turning over quickly and for better-than-market prices).

What about credit information? We're all concerned about our credit and a bit shocked to find how many hundreds of times a year that data is used to determine how we are marketed to. Today, there are some constraints on how this data is made available and accessed, but most privacy advocates believe them to be insufficient. The privacy advocates want complete control over the information to lie in the hands of the individual. But that leaves us with a problem: can credit-giving organizations hand out credit if the only information they can get their hands on is first approved by the applicant?

It would seem reasonable that if you didn't give permission to the credit agencies to have your information that they would have no knowledge of its existence, but that's ripe for abuse. Basically, somebody with bad credit would like identical to somebody with no credit.

However, another proposal is that the credit agencies would be allowed to know that information exists, but not get the details unless you released them. This sounds reasonable, but it is unlikely to be useful, because the only information most people wouldn't allow access to would be negative information. Thus, credit agencies wouldn't grant credit to people who don't practice full disclosure.

One interesting experiment is the recent move to secure medical information. Although not used for the same purposes, it is a similar kind of data and has much more personalized controls than it has in the past.