David R. Henderson  

NSA Surveillance: A Cost/Benefit Analysis

PRINT
Krugman on "Unproductive Finan... Apple's E-Books are Pro-Compet...
However, the reaction has continually been to expand the enterprise, searching for the needle by adding more and more hay. Far overdue are extensive openly published studies that rationally evaluate homeland-security expenditures.

The NSA's formerly secret surveillance programs have been part of the expansionary process. If they have done little to prevent terrorist attacks in the United States, and if we are now having what President Obama has characterized as a "healthy" debate about the programs, it seems reasonable to suggest that the debaters should at least be supplied with information about how much the programs cost.

Knowing the cost would scarcely help the terrorists. It might, however, amaze American taxpayers. Perhaps that's another reason the programs have been kept secret.


These are the closing paragraphs of John Mueller and Mark G. Stewart, "3 Questions about NSA Surveillance," The Chronicle of Higher Education, June 13, 2013. The whole piece is worth reading.

One other key paragraph:

It is difficult to see how earlier exposure of the programs' existence would have aided terrorists, who have known at least since the 1990s that U.S. intelligence was searching communications worldwide to track them down. It is possible, however, that the secrecy of the programs stems from the Obama administration's fear that public awareness of "modest encroachments" on privacy would make further efforts to encroach more difficult.

Basically, the gist is that the benefits of NSA surveillance have been very small and the costs have been quite large. Moreover, Mueller and Stewart bias the result against their own conclusion by understating the costs: they completely leave out the cost of NSA violating our privacy.


Comments and Sharing





COMMENTS (12 to date)
rapscallion writes:

The data nest egg the NSA is sitting on must be worth billions. Surely it could be put to additional productive uses, like helping people defend themselves in court.

Finch writes:

Leaving the morality and government power questions aside, and they are obviously huge and I don't mean to dismiss them, why would one think this program was particularly expensive?

Compared to satellites, spy-planes, and commandos, surely a billion or so on computer spying is pocket change? One success, say if the system had stopped the Boston attacks, would have been ample financial justification. So I don't think that part of the argument works.

David R. Henderson writes:

@Finch,
Did you read their whole piece? They didn't appear to have stopped anything approaching the severity of the Boston attacks. Which makes sense when the needles are already hard to find and they add more hay.

Hazel Meade writes:

My understanding that that the NSA is building an enormous data center in Utah to house this whole thing. Insane amounts of data storage capacity.

It would be nice to know how much that is costing, and get some idea why they need the vast amounts of storage capacity that are planned.

I vaguely remember back in 2002-2003 the uproar about "Total Information Awareness". That was ultimately quashed, but given the latest revelations it appears that they effectively have been doing it anyway, only calling it something else and keeping it secret.

Finch writes:

You mean the NSA didn't tell us all the strengths and weaknesses of their secret spying machine, even after some details got leaked? I'm shocked, shocked I tell you!

Further, David, the authors don't know the but-for world. It's like critiquing airport security. How do you know what would have happened without it? All you can reasonably say with public information is that this is cheap, and it looks like it helped a little in some public cases, but we probably don't know about much of the interesting detail that would be necessary to fairly judge its efficacy.

The benefit of stopping terrorist attacks is extremely high. If this were just about money it wouldn't worry me so much as there are vastly more wasteful things going on. But living with ubiquitous law-enforcement is deeply troubling.

The needles and hay thing makes a nice turn of phrase, but when I want to find something obscure on the internet, I turn to Google and it works pretty well. Google spends a lot of money adding hay to good effect. If I put on my evil dictator hat, this sort of technical means would be an obvious smart path to take.

Aaron Zierman writes:

Supporters of this surveillance program (and many other programs) prefer not to dwell on something like cost. They typically fall back on the "if it saves just one life, how can you be against it?" argument. Which implies, of course, that those concerned with cost care nothing about the lives of people.

Finch writes:

> Supporters of this surveillance program

For the record, I'm not a supporter of this program. I just think the cost argument is very weak, and the efficacy argument is hard to make with public information.

I'd appreciate a serious public think on what state/police powers should look like in the future. An analysis that was technologically aware.

Hazel Meade writes:

@Aaron, on the converse side, you might point out the cost to human beings that may occur if the surveillance power were to be abused (which it surely would be). Costs ich would come in the form of people being imprisoned unjustly, or persecuted for acts of political dissent, or simply because other, more powerful political actors find it convenient to get certain people out of the way for whatever reason (and I could cite the treatment of illegal immigrants as an example of what can happen to people).

Trouble is there is a whole cadre of people who will respond to those charges with cries of "cynicism" and "paranoia", as if the notion that the government might abuse it's power is some fantasy that libertarians just made up in the last year or two. (It could never happen here!)

Evil libertarians, always trying to subvert our faith in authority. Why are they so mean?

Aaron Zierman writes:

@ Hazel Meade

Absolutely. The costs are not only monetary in form. The difficulty is quantifying precisely what the costs are.

@ Finch

I was certainly not trying to imply that you were a supporter of this program. I think you make some good points about the trouble of trying to evaluate the monetary costs. I think it becomes even more difficult when we bring in the total potential costs. But I disagree with you in that I believe monetary cost still has value in the discussion, albeit a difficult cost to evaluate.

Finch writes:

> I was certainly not trying to imply that you were
> a supporter of this program.

Thanks Aaron; I wasn't sure and I realized I could have been read that way. When thinking about costs and cost-effectiveness, I was thinking of cash the government would have to pay for the program, not the broad effect on society. For example, this will probably make it modestly harder for US tech firms to do business overseas(*). That's a negative outcome, but it wasn't the kind of cost I was talking about. When I called for more discussion I was thinking about these second-order effects.

(*) Although this is kind of silly, since other countries surely have analogous programs. For example, there's been a long-standing rumor that the US and the UK have a cross-spying agreement on electronic communication.

MingoV writes:

There are "benefits" to the NSA spying on US residents that were not mentioned:

Sharing information with other law enforcement agencies so they can select "suspects" for additional spying.

Sharing information related to finance with other federal agencies such as the IRS, SBA, FHA, Dept. of Education (student loans), etc.

Sharing information related to political affiliation with key administration officials such as the president and select members of Congress.

Sharing information that could damage the image of politicians who are not democrats. (This, of course, will change when a republican is president.)


The above benefits have a value that almost certainly exceeds that of detecting terrorism.

Delphin writes:
extensive openly published studies that rationally evaluate homeland-security expenditures.
Which you won't get. Not from the NSA. When the NSA "published" articles in crytography years ago for conferences they would publish only a one sentence summaries. The details could not be peer-reviewed.
Comments for this entry have been closed
Return to top