November 23, 2017

Could Too Much Transparency Lead to Sunburn?

On Tuesday, the Houston Chronicle published a story about the salaries of local government employees. Headlined “Understaffing costs Houston taxpayers $150 million in overtime,” it was in many respects a typical piece of local “enterprise” journalism, where reporters go out and dig up information that the public might not already be aware is newsworthy. The story highlighted short staffing in the police department, which has too few workers for all the protection it is required to provide the citizens of Houston.

The print story used summaries and cited a few outliers, like a police sergeant who earned $95,000 in overtime. But the reporters had much more data: using Texas’s strong Public Information Act, they obtained electronic payroll data on 81,000 local government employees—essentially the entire workforce. Rather than keep this larger data set to themselves, as they might have done in a pre-Internet era, they posted the whole thing online. The notes to the database say that the Chronicle obtained even more information than it displays, and that before republishing the data, the newspaper “lumped together” what it obliquely descibes as “wellness and termination pay” into each employee’s reported base salary.

In a related blog post, Chronicle staffer Matt Stiles writes:

The editors understand this might be controversial. But this information already is available to anyone who wants to see it. We’re only compiling it in a central location, and following a trend at other news organizations publishing databases. We hope readers will find the information interesting, and, even better, perhaps spot some anomalies we’ve missed.

The value proposition here seems plausible: Among the 81,000 payroll records that have just been published, there very probably are news stories of legitimate public interest, waiting to be uncovered. Moreover (given that the Chronicle, like everyone else in the news business, is losing staff) it’s likely that crowdsourcing the analysis of this data will uncover things the reporting staff would have missed.

But it also seems likely that this release of data, by making it overwhelmingly convenient to unearth the salary of any government worker in Houston, will have a raft of side effects—where by “side” I mean that they weren’t intended by the Chronicle. For example, it’s now easy as pie for any nonprofit that raises funds from public employees in Houston to get a sense of the income of their prospects. Comparing other known data, such as approximate home values or other visible spending patterns, with information about salary can allow inferences about other sources of income. In fact, you might argue that this method—researching and linking the home value for every real estate transaction related to a city worker, and combining this data with salary information—would be an extraordinary screening mechanism for possible corruption, since those who buy above what their salary would suggest they should be able to afford must have additional income, and corruption is presumably one major reason why (generally low-paid) government workers are sometimes able to live beyond their apparent means.

More generally, it seems like there is a new world of possible synergies opened up by the wide release of this information. We almost certainly haven’t thought of all the consequences that will turn out, in retrospect, to be serious.

Houston isn’t the first place to try this—it turns out that the salaries of faculties at state schools are often quietly available for download as well, for example—but it seems to highlight a real problem. It may be good for the salaries of all public employees to be a click away, but the laws that make this possible generally weren’t passed in the last ten years, and therefore weren’t drafted with the web in mind. The legislative intent reflected in most of our current statutes, when a piece of information is statutorily required to be publicly available, is that citizens should be able to get the information by obtaining, filling out, and mailing a form, or by making a trip to a particular courthouse or library. Those small obstacles made a big difference, as their recent removal reveals: Information that you used to need a good reason to justify the cost of obtaining is now worth retrieving for the merest whim, on the off chance that it might be useful or interesting. And massive projects that require lots of retrieval, which used to be entirely impractical, can now make sense in light of any of a wide and growing range of possible motivations.

Put another way: As technology evolves, the same public information laws create novel and in some cases previously unimaginable levels of transparency. In many cases, particularly those related to the conduct of top public officials, this seems to be a clearly good thing. In others, particularly those related to people who are not public figures, it may be more of a mixed blessing or even an outright problem. I’m reminded of the “candidates” of ancient Rome—the Latin word candidatus literally means “clothed in white robes,” which would-be officeholders wore to symbolize the purity and fitness for office they claimed to possess. By putting themselves up for public office, they invited their fellow citizens to hold them to higher standards. This logic still runs strong today—for example, under the Supreme Court’s Sullivan precedent, public figures face a heightened burden if they try to sue the press for libel after critical coverage.

I worry that some kinds of progress in information technology are depleting a kind of civic ozone layer. The policy solutions here aren’t obvious—one shudders to think of a government office with the power to foreclose new, unforeseen transparencies—but it at least seems like something that legislators and their staffs ought to keep an eye on.

Comments

  1. Don Munsil says:

    I just recently learned that the tax returns of all citizens are public records (and indeed are published in a book every year) in several Scandinavian countries. IIRC, it was Finland, Norway, and Sweden.

    At first I thought this was crazy, but the more I think about it, it seems like it probably just makes public the exact details of what people can already figure out to a fair degree of accuracy from information that’s already readily available.

    Certainly these countries have been doing this for years, and folks don’t seem to mind. In general, I think we may be too obsessed with privacy. Why would it matter if marketers could get all our income data if they weren’t allowed to make bothersome marketing calls? And wouldn’t I rather get advertising targeted at my demographic?

    In general, I’m more concerned with reining in practices I don’t like than trying to hide my information. As we all know, trying to restrict the copying of information is doomed (long-term) to failure.

    I do think it’s unfair to have unequal disclosure. Let’s either say that income is private across the board, or it’s public across the board.

  2. This is really just a specific example of a larger issue, namely that just because we can collect, collate, and search ever larger databases doesn’t mean that these databases should be made.

    Whether we should do the things that technology allows us to do is something that I don’t think is discussed enough in general and in particular when databases of information about people are concerned.

    Perhaps I sound like a luddite, but I think those small obstacles are more important than we generally think.

  3. This isn’t all that surprising. Anectdotally, a couple of local newspapers (Boston and Jacksonville, FL) that I’ve seen publish the same kind of data online (though I don’t know exactly at which level of employee the information stops). Don’t have the url’s at hand, but I know that it is out there, to some extent.

  4. As an employee at a state university where the student newspaper has a standing FOIA request for all staff and faculty salaries and publishes them every year, I’ve experienced this odd situation personally. On the one hand, it was a little unnerving to discover that my salary was publicly available on the open web. On the other hand, there are ways this openness benefits the people who work at the university. It makes it very easy to discover systematic pay disparities between men and women, and to remedy them (which has happened). It helps workers who are just starting out get a sense of what kinds of pay increases they can expect to receive, and when. It provides helpful context for people asking for raises. I’ve come to believe that salary transparency at my university, while it may have its drawbacks, is as beneficial for the employees as it is for the general public.

  5. To Don Munsil:
    I cannot say much about Finland and Norway, but I can say that in Sweden the tax return you file is available to the public. Not all of it, but you can see the first page of it by just asking. The first page contains the sums arrived at in other parts of the form, so you don’t get to know how much money was made by working or investing, you just get to know how much money was made in total.

    When it comes to people employed by the cities, if you want to learn about their wages you need do no more than call city hall and ask: it’s a fact openly accessible to the public. I don’t think city hall is even entitled to asking why you want to know.

    The basic assumption about official papers (and information) is that it is all openly available to the public. The only cases when papers/information are not available are, as far as I know, when the informations propagation would damage the country (think military secrets) or when the information would damage the victim of a crime (think court records). Please note that the secrecy of court records only protects the victim of a crime, not the criminal.

    I might be wrong on some small details in the above text, and give no guarantee as to it’s corectness. I do, however, live in Sweden.

  6. Does privacy mean nothing? Let’s see Cindy McCain’s records out in the open – for which a very good case does exist. How about Rush Limbaugh’s? Just where did all that drug money go.

  7. I think RJD has an important point: what we’re in is an era of partial, erratic transparency. Every point you make about supposedly ferreting out corruption (maybe the people in question have spouses with better-paying jobs, or generous relatives, or have been saving instead of spending for the past 20 years) applies to to everyone who works for an enterprise that does business with the government.Especially in the era of privatization, the question of whether one’s paycheck comes directly from the government should hardly be the thing that determines whether people undergo proper scrutiny. And it’s not just house purchases, let’s buy lists of known-good prospects from direct-marketing firms and cross-reference those too.

    I’m sure that once the service is well-established even private companies that do no government business would love to get in on the act and deliver their payroll files to a third party so that their employees can be vetted for possible corruption.

    But whatever you do, don’t extend this kind of oversight to anyone in the upper brackets.

  8. In order to judge the value of privacy, it is first necessary to ask what purpose it serves. The following article is a brief overview of spite and fairness.

    http://www.newscientist.com/channel/being-human/brain/dn10239-sense-of-justice-discovered-in-the-brain.html

    By hiding the wages paid to others, it is more difficult to detect unfairness, thus blocking people’s spiteful enforcement of what they consider “fair”. This brings up the deeper question of why such a brain system evolved in the first place, what purpose is served and the relevance to the modern world.

    There’s a further purpose to privacy, which prevents a major group from victimizing minority groups by allowing people to hide aspects of their lifestyle. It is easy to find examples of minority groups being scapegoated as convenient political targets, generally on racial or religious grounds. We would not like a buildup of intolerance to result in a society without diversity.

    Of course, by far the most popular purpose for privacy is for people with power to find methods to secretly abuse the power entrusted to them, in order to bolster their own self-interest. Since these are generally the people who get to define the rules of the system, so the system tends to be constructed with the interests of such people placed ahead of everyone else.

    For example, money is a sterilizing agent — a million dollars gained through hard work is equal to a million dollars gained through theft. Money is one dimensional, you can have more of it or less of it but not different types of it. On the other hand, actions are multi-dimensional. If you could judge someone by their past actions then you might judge very differently to being able to judge only by their ability to pay, but you are not given this option because that history is considered private. We know that billions of dollars every year change hands in illegal drug sales, yet this drug money rapidly merges back into the rest of the economy without a trace and everyone accepts this as perfectly reasonable.

  9. Speaking as an employee of a public library, I don’t see what the big deal is — I’m paid primarily through tax dollars, and if ever there were something that should be public, it’s how the government spends money.

    Personally, I want more openness in government — the tendency is for government figures to hide behind secrecy when the real problem is that they screwed up and don’t want to have to deal with the repercussions of that failure. I think that public sector workers should understand that working for the government means more than just having a different entity paying the check.

    Privacy FROM the government is an entirely different issue, of course.

  10. “kfqsuw adbcugf mkcz”

    Some kind of cryptogram? It doesn’t look like any language I’m familiar with. That last word has no vowels, in particular.

  11. “There’s a further purpose to privacy, which prevents a major group from victimizing minority groups”

    And this, by itself, justifies privacy and the efforts taken to secure it. Also, the need to keep cryptographic secrets to be able to prove (when needed) your identity.

  12. More transparency leads to freedom……………this is good for America

  13. Too much transparency can put minority groups in danger. Sometimes even physical danger. Imagine every gay suddenly outed, in the current climate. In some far eastern country they just arrested some guy simply because he had gay sex — it didn’t say anything about it being nonconsensual; apparently it was simply because it wasn’t straight. Your transparency, if it suddenly developed overnight, would result in at least ten thousand lynchings and Christ alone knows how many unjust imprisonments, beatings, and other attacks by sundown the following day.

    Perhaps your ideas could work in a society that has matured enough to be very tolerant of harmless variations, to be intolerant of intolerance itself (except for intolerance towards intolerance…), and to no longer enforce any laws against victimless crimes. But I don’t know of such a society here on Earth.

  14. @Spudz; Slightly off topic: You are aware that the main difference between “some far eastern country” and most US states is that they actually enforced the law?

  15. I worked for a company that did a salary survey, and then published the salary ranges for every job title in the 50-person department. Since I was the only person in the specific role in the department, everyone knew my (new) salary range. This worked the other direction, too. It was actually pretty liberating.

    If you’re a government employee concerned about ‘nosy neighbors’ snooping on your salary information, then why not impose the requirement that they have to make their (accurate) salary available to see yours. With proper database design and access controls, that could work.

    The only one who really benefits from the secrecy is the employer who wants to pay some employees in the same job different amounts, by swearing them to keep their salary secret.

  16. Tarkeel: yes, there are bad laws on the books in some US states. It’s wise that they don’t enforce them. It’s ridiculous that they don’t remove them completely from the books.

    None of which alters my point about the likely consequences if suddenly privacy completely disappeared tomorrow.

  17. As all the comments have indicated, privacy is deeply contextual. What is the value of exposing or recording or publishing information? The value of Amazon and Google tracking all my actions/behaviors is for private profit. That is either a direct exchange of value or them effectively leveraging peer production. The risk there is governmental access, or resell of information for secondary purposes which could then cause residual harm as it ends up in a database of attitudes and behaviors accessible by government and usable for private hiring agents to discriminate.

    The exposed information in this case enables comparisons directly between male and female salaries, for example. It shows the rhetorical point that not hiring and failing to staff costs government/taxpayers money. Linking medical data; however, or posting dates of approved vacation time that could indicate medical issues is something that I would strongly oppose. There is no value in exposing a private medical condition to strangers.

    Americans secrecy about our salaries only helps the powerful and individual organizations discriminate, primarily against women but also men of color.

    In summary this information is macro (what I make not details on what I spend), enables important identification of violations of civil rights, and serves a democratic function. I support it. Here I put my money where my post is:
    http://www.heraldtimesonline.com/htoxtras/salaries.php

  18. pafkoseora says:

    Hello my friend, your site is very good! http://ioownbouojjki.com

  19. What’s with these last two posts? The first seems to be some sort of cryptogram (the vowel frequency and placement is wrong for it to be any known natural language) and the second is, while friendly, not very specific to the topic…

  20. Another cryptogram?