Theres an interesting discussion going on between Juan Volokh, Mark Kleiman, and Dave Kopel about ballistic fingerprinting, which I discussed back here. My proposal was to take ballistic data from guns, and store them in a series of private registries, who would hold the data and registration data in such a way that a) only a limited set of queries would be answered, and b) the registry as a whole would be volatile, and could be destroyed by the registrar in the event of an effort by the government to take it (which destruction would be explicitly permitted in the enabling legislation).
My suggestion was way too complex. (Although it raises other interesting possibilities)
Mark Kleiman proposes a simpler (although somewhat less effective) system, whereby spent brass (and possibly bullets) would be filed and imaged in a database and stored solely against the make, model, and serial number of the gun.
The manufacturers existing systems can track the gun to the wholesaler, whose records would reveal the retailer, whose records would then link to a purchaser, which would take a match at least as far as the original new purchaser of the gun.
I have one argument for, and two against, this proposal.
The argument for is:
1) It will help solve some certain number of crimes that would otherwise not be solved.
The arguments against are:
1) It will cost money which would be better spent on other crime-prevention or crime-solving resources;
2) It wont work very well because
a) the image matching technology isnt very good, the markings and characteristics of each gun change over time (or are easily changed), and so there will be an extremely low hit rate;
b) it will only be applied to new guns sold after the effective date (otherwise it begins to look like registration)which means it will apply to only a small fraction of the guns in circulation.
So on a cost-benefit basis, I wonder how effective it would really be. This ought to be (roughly) calculable, as an example: we assume that out of 500,000 crimes committed with guns, say 5,000 (made-up numbers alert!!) are committed with guns purchased in the last year. Of these, 1,500 are unsolved. Of these 50% involve spent brass or intact bullets, and we have a 20% chance of getting a match in the system, so we ought to be able to track 150 of these guns to their purchasers. Of these, 75% still own them, so we might get 120 hits on a system like this. One issue would be the number of hits we would get through simple good police work, so the additional hits would be, hypothetically, in the range of 60 80. Note that these numbers are entirely made up, and that Im simply trying to demonstrate the mechanism whereby we could cost a system like this.
Imagine that 1,000,000 guns are sold every year (thats less than I believe are actually sold, but sounds about right and Im not doing research tonight); it costs $10.00 to fingerprint, image, and store the spent brass and bullets (real cost is probably double); were talking about spending $10,000,000/year to help solve 70 crimes. Now that will go down every year (in ten years, well be spending $10,000,000 to solve 700 crimes). But on numbers like that, Id say the proposed plan is a bust. Given some time, Ill try and plug real numbers into this, and see how they come out (or you can do it and email the results to me!) but the question to the audience ought to be, how much do we spend on a program like this per possibly solved case?? $10,000?? $100,000??
There is a broader issue as well.
As I noted, On the other hand, there are a large group of people in this society who hate guns, and devoutly wish to make them go away
at least except for the ones they get to carry (see CA state Senator Beretta Perata) or their bodyguards carry (see Rosie). And these people are close to the levers of power, and it isnt hard to imagine that one day theyd get those levers, and use them to do whatever they could to take guns away from everyone who wasnt them. Theres a logical chain that goes from fingerprint new guns to fingerprint all guns on sale to fingerprint all guns and suddenly, like the mythical frog in a pot, were being boiled.
On one hand, Im convinced that the potential for this is very real.
On the other, I believe that the best way to resolve it is to politically defeat the hoplophobes (Col. Coopers the other Jeff Coopers term for those who are afraid of and wish to ban the private possession of weapons) completely and thoroughly.
And to do that, we (the gun-rights defending population) need to find a way to reach out to the big group in the middle. Some of them own guns; some of them are afraid of guns. One thing we need to do (and that I do all the time) is to stand ready to discuss the issue of violent crime using guns and what can be done to limit it. And having had those discussions, we need to stand ready to support measures that a) dont directly remove our rights or lead in that direction and b) have some hope of being effective (including being cost-effective).
I think that ballistic fingerprinting, as suggested by Kleiman, passes test a). Im not so sure about test b), but that ought to be something where some reasonable set of facts can be brought to light.
I have some homework to do, I guess. Anyone want to help?
[Addendum: Don’t forget to read Rob Lyman and Steven denBeste on these issues. I infer to a point they make … that freedom has a cost, and just because we’re seeing it doesn’t mean we should stop paying it … but I seldom make the argument as persuasively as they do.]
The “database can be destroyed” safeguard is least likely to work in the situation where it is most needed. (The govt in question announces that it wants the databases and will kill any administrators whose databases are unavailable.)
Also, the safeguard is just legislation, and legislation can be changed. And, we have relevant experience. For example, NYC promised that its registration scheme wouldn’t be used for confiscation. Yet, when NYC decided to ban certain guns, guess what happened. The second set of legislators helpfully pointed out that they’d made no such promise. It’s more important to realize that such promises aren’t binding.
Klieman seems to think that the 4473 system is not registration.
I have a different view. The relevant question is whether a mechanism is very helpful for confiscation.
All of the 4473s for out-of-biz gun stores are in ATF warehouses AND they’re usefully indexed. The US govt also has the ability to get 4473s (either individually or in mass) from gun stores.
Thus, the 4473 system is quite helpful for mass confiscation.
Klieman might respond that the relevant question is selective confiscation, say, confiscating guns from all members of . The 4473 system doesn’t give the govt a list that they can cross-check against the enemies list.
Ah, but as I have shown above, the govt can easily have such a list. It’s less trouble than creating the enemies list.
Klieman seems to think that using cases instead of bullets addresses technical objections.
One of the interesting things about cases is that whether or not they are available at a crime scene is something that the shooter decides. Shooters who use revolvers almost never leave cases, and those who use semi-autos can decide not to.
Also, it is just as easy to change the case markings. With the exception of revolvers, the case is inside the initial part of the barrel when fired, so in most cases, changing the barrel changes the one part that makes most of the marks on the case. (With revolvers, the cylinder can be readily changed.) The other parts are readily changed as well.
And, it’s easy enough to simply change the relevant surfaces of these parts so that the marks are different. When this is done, the database entry can not match cases from the gun.
I’m a little suspicious of the idea that guns’ characteristics change over time and can easily be altered. I’m sure that there’s some truth in this but I suspect that the degree has been exaggerated by second-amendment advocates. (Speculation).
Suppose you alter the bore of a gun barrel sufficiently to change its signature. Would that change the gun’s accuracy? It seems that if it was a good gun to begin with any change would be negative. (Speculation).
A system like this wouldn’t be a cure-all, especially initially. But it could be useful in many cases, the more so the longer it was in effect.
>> Suppose you alter the bore of a gun barrel sufficiently to change its signature. Would that change the gun’s accuracy?
No and no.
Very good parts are accurate to 0.001″ but most aren’t anywhere near that good. The marks are made by features that are far smaller. As a result, one can make completely obliterate these features and the gun is still within “as new” specs.
The forensic folks talk about recognizing matches when they know that the guns are the same “even after 1000” shots. Barrels easily last for 20-50x shots without any measurable change in accuracy. So, if someone is willing to sacrific half of his barrel’s life to thwart a match, he’s >10x beyond the best case for forensics.
And, even after a barrel is “measurably deficient”, it isn’t significantly deficient. Few people can shoot anywhere near as good as their gun, so a gun that is half as accurate is just as effective.
With respect to the parts that “scratch” cases, they can be modified by several times 0.001″ without affecting a thing. (These parts are, with one exception, no where near as precise as barrels to begin with.)
And then there’s the parts swapping issue.
>> But it could be useful in many cases, the more so the longer it was in effect.
Based on what evidence? It’s not like the idea is new and untried.
Also, if you’re going to argue “let’s try”, then surely it’s fair to ask what happened after other experiments failed. Until the advocates of “let’s try” get around to advocating the repeal of things that haven’t worked ….
>> I’m a little suspicious of the idea that guns’ characteristics change over time and can easily be altered. I’m sure that there’s some truth in this but I suspect that the degree has been exaggerated by second-amendment advocates.
And what is the basis for these suspicions?
I’ve heard gun controllers complain that certain things would work just find if only the pro-RKBA folks would cooperate. For example, “you know the difference, you just won’t tell” was a fairly typical response when pro-RKBA folks pointed out that “assault weapons” weren’t functionally distinct.
Besides the gradual (by use) or rapid (by deliberate action) changes to the recorded characteristics, there is the issue of 220 million guns already in civilian hands – those will get into the database only after they are used in a crime and leave evidence behind – unless all guns already sold are required to be entered into the database.
The cost issue is addressed by Clayton Cramer, at
http://www.claytoncramer.com/weblog/blogger.html.
It’s the first thing under “Thursday”.
The fact that the 4473’s already exist asa ‘de facto’ registration regime is one of the reasons I tend to be softer on these things than many of my friends; we’re not going from ‘nowhere to registration’, we’re going from ‘kinda registration’ to wherever…
A.L.
Here’s an e-mail that I sent to Mr. Kleiman, and copied to all parties involved:
Mr. Kleiman,
I linked to your site from The Volokh Conspiracy blog, and read your link to Dave Kopel’s article, and have some comments I hope you’ll indulge.
You wrote: “The crime-solving benefits of the proposed system are substantially greater than Kopel allows, and the primary disadvantage he cites — the creation of a nation gun registry — is a complete fantasy.”
Um, no – on both counts. Registration is a very real possibility. With each and every firearm law there’s always a “loophole” that requires a “next step” to close. I foresee several “loopholes” with a “ballistic fingerprint” database, foremost being that the serial numbers are not directly tied to a possessor, second that there is no required chain-of-ownership documentation (the “gunshow loophole” reborn,) and third that the guns currently in private hands are not “fingerprinted.” Each of these “loopholes” points towards registration. A “fantasy” it is not. It’s another step down Eugene Volokh’s “slippery slope.”
You wrote: “We don’t have to speculate about how well ballistic imaging works; it’s being used now on shell casings recovered from crime scenes, and there have been thousands of ‘hits.'”
Well, yes. However ballistic fingerprinting is currently run on known or suspected crime guns against a relatively small database of crime scene evidence. I invite you to read the paper “Technical Evaluation: Feasibility of a Ballistics Imaging Database for all New Handgun Sales” produced by the California Department of Justice in October of last year. It is available here: http://www.nssf.org/PDF/CA_study.pdf Please study carefully the technical difficulties that expanding the database simply to include all NEW handguns would cause, the problems currently involved in matching “ballistic fingerprints” when the ammunition used for the “fingerprint” is of a different manufacture than the one found in evidence, and so on. Note also the commentary about changes in the markings on both the cartridge case and bullets due simply to use.
You wrote: “John Allen Muhammed bought the rifle that was found in his truck (actually, I thought it was found in his car), and which has been ballistically linked to the DC-area sniper shootings, new, from a gun store, in his own name. Had we been collecting test-fire shell casings, we might have had his name and driver’s license photo (available for law enforcement purposes from most state motor vehicle registries) ten days ago. That might have saved some lives.”
Yes, it might (though I believe the match came from comparing bullets, not cases.) However, it raises some other questions. How many rounds had Muhammed fired through his AR-15 between the time he purchased it and the time he began committing murder with it? Would that have reduced the probability of a ballistic match? What ammunition would be used for the initial “ballistic fingerprint” and what ammunition did Muhammed use to kill his victims? Would that also have affected the probability of a ballistic match?
One crucial question: What if Muhammed had changed the upper receiver on his AR-15? If you were not aware, like an automatic handgun the AR-15 has two major sub-assemblies; the upper and lower receivers. The LOWER receiver is legally the “gun” – the part that carries the serial number and that is tracked by the BATF from manufacture to dealer. I had a custom AR-15 built for me on a new Bushmaster “lower” that I purchased “stripped” (it contained no parts – just the machined lower receiver with a serial number stamped on it.) The lower cost approximately $150. How do you “ballistic fingerprint” a stripped receiver? I sent the stripped lower to a custom builder who then assembled a rifle for me, and shipped it back to me as a complete firearm. That added about $1,500. I can, at any time, mail-order a new complete “upper receiver” to put on my assembled lower receiver for between $400 and $1,000 depending on its configuration. (See this link for an example: http://www.midwayusa.com/rewriteaproduct/325576 ) The upper is not “a gun” according to the BATF. No background check is required. The two pieces are held together with two pins designed to be easily pushed out with a cartridge. This is one of the features of the AR-15/M-16 that make it simple to clean. The upper receiver is the part that leaves the markings on bullets and casings, not the lower. I can do precisely the same thing for my Kimber .45 pistol – change the barrel and the slide, and the “ballistic fingerprint” will no longer match the serial-numbered frame.
Ballistic comparisons of known or suspected crime guns against a limited database of known crime-scene evidence is a useful procedure. The firearm tested is usually contemporary with the evidence, and it is unlikely that any major modifications have been made to it. By mandating a “ballistic fingerprint” database, it would alert the criminally inclined to take the remarkably simple steps necessary to render the database useless. Building a database of all new (or simply all) firearms would be a horrendous waste of time, money, and manpower – another “feel-good” gun-control measure that (like the majority of those implemented to date) have proven ineffective at best, counterproductive at worst. It would be just another “next step” down the slippery slope.
Zizka–
A process called “lapping” is a common technique for making a gun shoot MORE accurately. In other words, I can eliminate the signature AND get a better rifle.
Now, proper lapping takes considerable time and skill, so I doubt many criminals will use it. But steel wool and a cleaning kit will do a “good enough” job that doesn’t destroy the gun.
On the other hand, the huge majority of gun crimes are committed with handguns, and I can swap the barrel out of a Glock or Baretta in under 30 seconds.
As for the wear-and-tear question, it would be the easiest thing in the world to check it out–it would cost a news network maybe $10-$20K, for a crude John Stossel piece, but they haven’t done it yet. Frankly, I don’t really know how serious the problem is myself, and I can’t afford to test it.