Wednesday, June 12, 2013

Tree Don't Care What A Little Bird Sings

. Wednesday, June 12, 2013



I have not read much of Robert Fogel's work, not much at all, but I may need to read more of it. A Fine Theorem, one of the more under-appreciated blogs, has a summary of Fogel's Without Consent or Contract. Here's part of it:

... the paradox rests on the widely held assumption that technological efficiency is inherently good. It is this beguiling assumption that is false and, when applied to [American] slavery, insidious.”  

Roughly, it was political change alone, not economic change, which could have led to the end of slavery in America. The plantation system was, in fact, a fairly efficient system in the economic sense, and was not in danger of petering out on its own accord.
Here's the rest.

There are multiple views of the politics of technology. (Technology is, at its core, information aggregation.) One says that technology is liberating. Another says that technology is enslaving. Another says that technology is fueled by the state for purposes of control. (Oddly, skeptics of markets often make the first point of that point without understanding that the second point is the corollary.) Technology can destabilize the political equilibrium (but does that only apply if it goes in one direction? I doubt it). It's worth googling a bit for the views of Farrell, Drezner, and Lynch on this. It's worth noting that modern authoritarian regimes try to get to the technological frontier as rapidly as possible but they tend to have a tough time managing it. Francis Spufford's Red Plenty is on sale at Amazon right now, if you don't mind probably giving some of your metadata to the NSA.

Sarah Jaffe (on Twitter) asked for a political economy of the surveillance state. (Here's a short take, not very good.) I haven't got the time or background knowledge to build a real model, but if I was going to I'd start with Tilly and Scott and Weber at the foundation and ask what purpose this really serves. Knowledge is power, is it not? Power is needed for protection (in the Tillian sense), is it not? After that I'd go to Orwell like everyone already is, but not the dystopian cliches. Remember in 1984 that Winston Smith was pretty much the only one in society bothered by Big Brother. (Probably not, if you've read your Timur Kuran, but as far as Smith could tell he nearly enough was.) Everybody else just got on with it. The proles sang their songs and read their magazines. Sure, Julia was a bit inconvenienced by the whole thing, but it's not like she really had principles.

Now think about Havel. Now think about samizdat. Is information so easily controllable? Can the state not oppress on the basis of allegation, innuendo, or missing data? Can the citizenry not resist simply by living? Does the state need all information to "keep the locals in line" or just a vague threat -- the vaguer the better? Corey Robin addresses this and gives a precis of his book on the politics of fear. Stalin didn't have Bukharin's metadata... just the ability to credibly say "we know where your kids are". That hasn't changed. Yglesias is right: the biggest thing to fear from the surveillance state isn't the state, per se. But that's a micro story, and micro stories can dictate macro policies.

The U.S. public is not concerned about this. To the extent they are it's for partisan reasons, not out of principle. Note that this is not new. Note that, so far, it appears that these programs are legal at least in broad terms. Intellectuals are more concerned that the median pollee, as they should be, since they are much more likely to be targeted than a randomly-selected person. (If I was Glenn Greenwald I'd go back to snail mail and pay phones for a good long while.) But so? Democratic politics does not guarantee puppies and roses. As we debate whether or not this is constitutional we should remember that James Buchanon's insights do not only apply to economic policy. We should also remember that politicians and celebrities have been subject to heavier levels of scrutiny than this for as long as there has been human society.

Data, even metadata, can be used for ill. (Or good, as the case may be, since the 21st century version of Paul Revere is probably someone Healy wouldn't meet for a beer at Ye Olde Tavern. Possibly this isn't what Healy's driving at.) But let's not get carried away. The U.S. government is sophisticated in many ways, but this program has only $20mn in funding. Let's say they spend $5mn of that on high-powered computers (that's probably less than what the supercomputer I ran a bunch of my dissertation on cost), and the rest on twenty-somethings making $200k/year each (as Snowden apparently did). That's 75 guys trying to make sense of the 2.5 quintillion bytes of data created each day. Good luck with that. (No I don't believe only $20mn was funneled into this. Not for a moment do I believe that. But I'm not sure how much $20bn could really do absent some good old fashioned police work.)

So after you've read the Spufford (or even before) you might want to read some of the discussion at Crooked Timber on the book. See especially this wonderclass by Shalizi which has as much to say about social science theory and methods as it does about historical political systems or the contemporary political economy of the surveillance state or novels. The key question is Shalizi's first one: what is being optimized?

Then recall that Hayek's slippery slope is a logical fallacy to which the historical record is not kind. Should we be less concerned? Probably depends on how concerned you were in the first place... anonymity is a myth.

Remember too that the government oppresses and kills and makes terrible decisions when it doesn't have good intelligence. Given that, is the expected utility of (American or other) society better or worse with PRISM or without it? Apparently this program stopped one or more attacks at the London Olympics. What would the cost of those attacks have been? Was preventing them worth $20mn dollars plus some false positives? (The TSA spends $6.5 billion a year and probably gets almost nothing for it.) Could PRISM have stopped Nidal Hasan had it been better-implemented? If it could have, would it be worth it? We are quite literally behind the veil of ignorance at the moment (just a bit less in the wake of Snowden's leaks), but if we take engaged citzenry to be a desirable normative end in itself we need to put our Bayesian caps on now and start updating our priors.

What tail event has a greater probability: that this program is abused in such a way that it devastates liberal society, or that it prevents a significant attack the fallout from which would devastate the same society?

In the end the biggest repercussions of NSA spying might be felt in the US-EU trade negotiations.

Nevertheless, I oppose PRISM and related programs very strongly. I do so because I am not risk-averse.

I believe this is the most Cowen-esque thing I've ever written. I also believe that every link in this post is worth clicking on.

0 comments:

Tree Don't Care What A Little Bird Sings
 

Share it

PageRank

SiteMeter

Technorati

Add to Technorati Favorites