Articles, Blog

L&CP Symposium 2019: Gun Rights & Regulation Outside the Home | Panel 3

afternoon, everybody. I’m Phil Cook. I’m in the public policy
school across the street here. Welcome, everybody,
to the section that is advertised as on
Empirical Work in Theory. In my world, there’s
nothing else. So I’m not sure– this is everything, right
here, for the next hour in the quarter. We’re going to start with Nick
Johnson of Fordham University Law School, a little bit
different than the sequence is in the program, but apparently
it’s not contractual even here in law school,
so we’ll start with Nick and then go on to John Pepper
who is a professor of economics and also public policy at
the University of Virginia. And then finally
with John Donohue who’s a professor of law
at Stanford University. We’ve got an all-star
lineup and 20 minutes each. Here we go. NICK JOHNSON: Thank you. So, my effort here is
to combine something that I’ve done before
and talked about with– a place where I’m not
sure that it operates. The title of the paper is
“Defiance and Concealed Carry”. And so in prior work, I
posited that two phenomena– things that I called
the Remainder Problem and the Defiance Impulse– operated to render
supply-side gun control– which I define as
sort of sweeping bans on broad classes of technology– rendered it
unworkable in the US. And I argued that the
internationally demonstrated tendency toward
defiance in combination with the exceptional
number of firearms that we have in
the United States would hinder the effectiveness
of supply-side gun control in a variety of
ways make things worse. So this paper considers
defiance in the context of the slightly different
problem of concealed carry restrictions and
I hope to use it as a way of illuminating
the issues that arise in the
context of concealed carry with a
particular focus on how defiance by political
subdivisions affects this general
notion of enforceability of concealed carry
restrictions and a subsidiary focus on the impact
of defiance and race. So in the first
part of the paper, I talk about the
defiance phenomenon, which in earlier work
I tracked, as I said, based on international data. And so there’s a nice
synopsis of what I focused on in the prior paper done by
the international small arms survey– this is 2007– and utilizing data
from 77, 78 countries. They ended up compiling a list
of basically massive defiance of restrictions on
either gun supply or defiance of registration. And what it seemed to
me was occurring there, was that we could not
explain the defiance on the basis of simply
the view that criminals have an inelastic demand
curve for firearms. It seemed as if a large
proportion of ordinary people– previously law-abiding
people– were also defying certain kinds
of gun regulations. So extrapolating from what
happened internationally, I posited that the
United States should– if we ended up in a scenario
where, either at the state or at the federal level,
there was an attempt to implement very
significant supply controls– that the United
States should face similar sorts of resistance. In fact, I posited
that we would see greater levels of resistance
given our unparalleled gun culture. We’ve got a
constitutional protection of arms that is unparalleled. Americans own half
the private firearms in the country, and Americans– more? JOHN DONOHUE: In the world. NICK JOHNSON: Excuse me. In the world. –own half the private
firearms in the world. And there is this
expectation that Americans would resist firearms
registration and supply restrictions in a way
that might generate problems that are I think
understudied in this context. Now, for the
particular experiments, in the early stages, the first
generation of assault weapons legislation in a
variety of states, Jim Jacobs found that
the compliance rates were in the single digits. So 3% in New Jersey,
1% in California. More recently we’ve
gotten other tests. So in December of
2018, New Jersey banned all ammunition magazines
that hold more than 10 rounds. Estimates extrapolated
from sales records put the general number
at around 2 million. The reporting suggests
that in New Jersey that no one has surrendered
a high capacity magazine. Now, like other technology
bans at the state level, this is not a clean
test of defiance. So it was possible for
alternative compliance to occur. People could modify
the magazines and render them legal. Also, one could simply sell
the magazine out of state and still be in
compliance with the law. And it’s possible that every
one of the New Jersey owners complied with the
law in that fashion. But there also is lots
of rhetoric surrounding this law with
observers skeptical given the kind of
defiance rhetoric that activists have been
utilizing proclaiming their intention to defy. Another current example, New
York’s assault weapons ban– dubbed the SAFE Act– generates similar
sorts of results. So the number of actual– the estimated number–
of assault weapons that would be required to be
registered under the SAFE Act is about a million. It looks as if about 50,000
currently have been registered. Now again, registration
was not the only option. It’s possible that alternative
compliance occurred. One of the commentaries, though,
surrounding the SAFE Act struck me as significant– and I’ll
talk about it in a little more detail in a bit– there’s a Forbes
Magazine article filed by a correspondent
from Dutchess County who claims to be a gun
owner, and he says, “There’s profound
social stigma among gun owners against registering
these guns with the government.” And then, more
importantly, he says, “They are quick to tell you that
many municipalities and County Sheriff’s departments have
reported that they will not enforce the SAFE Act.” So this gives us a hint of a
variation on the defiance theme that I’ll talk about in
a little more detail. That is, defiance that is
endorsed by local government or political subdivisions,
county commissioners, et cetera. I think that has a significant
impact on the way in which we expect defiance might
play out, and we’ll talk about that in a second. Another example, 2013. Connecticut banned
semiautomatic rifles with at least one
military characteristic. About 50,000 were– so it
was not actually a ban, it was possible
to register those that were essentially
grandfathered– about 50,000 guns were registered. The estimates from
national sales data put the number of unregistered
guns at about 350,000. Then there was
also a restriction on high capacity magazines. Those magazines had to be
what they called “declared.” About 37,000 of the
magazines were declared. Connecticut’s nonpartisan
Office of Legislative Research concluded that there were
about two million magazines in the state. Again, the possibility
of alternative compliance makes this not a clean test. Last example with
the state data. California’s preliminary
reporting shows roughly 3% compliance with the state’s most
recent technology restriction on rifles with
so-called bullet buttons and we can talk
about bullet buttons in the context of the
technological or challenges of tracking the technological
innovations in a second. The examples where–
other examples where the political system gives
formal voice to the defiance impulse occur. For example, in the
emergence of what are called Second Amendment sanctuary
cities or counties. These are places where the
same strategies deployed by immigration sanctuary
cities are being utilized. We find other examples
just in the declarations of the chief law enforcement
officers of various counties. So local sheriffs and
county commissioners in the state of
Washington, in Maryland, in Oregon, in Illinois, in
New Mexico, and Colorado have declared that they will
not enforce gun laws that have been enacted at
the state level, that are inconsistent with the views
they say of their constituents. At the state level, the
judicially invalidated Firearms Freedom Acts prefigured a
similar vein of defiance. Note these litigations
surrounding Firearms Freedom Acts in Montana have rendered
for now the model legislation ineffective and moot. But the legislation had
the impact essentially of nullifying the National
Firearms Act for firearms that were manufactured and
traded exclusively within the boundaries
of the particular state. And the National Firearms Act,
as you may or may not know, is the legislation that
controls the dissemination and possession of fully
automatic firearms and short-barreled rifles and
silencers and other variations of firearms technology
that many of us would think are among the most
problematic renditions of firearms technology. The Firearms Freedom
Act legislation by 2013 had been enacted by Montana,
Idaho, Wyoming, South Dakota, Utah, Arizona, Kansas,
Tennessee, and Alaska. So my claim is that
these phenomena are a significant version
of the defiance impulse. And my question,
then, is whether we have to think about this
impulse as we imagine the manner in which a
restrictive carry regime might be implemented– particularly
a restrictive carry regime that might occur after a change
in judicial philosophy. So we’re talking in a variety
of contexts about the legitimacy of concealed carry. We’ve got a recognition that
there is a sort of judicial agreement or
disagreement– excuse me– over the scope of the
constitutional right to carry outside the home. So imagine a scenario where
the judicial temperament shifts and we find that some
jurisdiction that has been under litigation
pressure forced to impose a kind of liberalized
concealed carry regime, reverts to a more
restrictive regime. So my question, essentially,
here is how will defiance– this impulse that
we have identified, if at all– how will it
impact that sort of move to or a reversion to a
more restrictive regime. The first thing that one
worries about in this context is basically measurement. So I’ve already
talked about how we’re getting these complicated
tests of defiance in the states where alternative
compliance is possible. That is, you can send
your banned technology out of the state– sell
it out of the state– and you’ve complied
in a way that’s not going to show up on the
state’s recording of people who have registered
their firearms or registered their magazines. Concealed carry is
different in the sense that there’s not
this complication of alternative compliance. The defiance is going to occur
within the jurisdiction that prohibits it, and it’s
going to occur immediately. But I believe concealed
carry is going to be harder for us to track. We don’t have a baseline. It’s going to be difficult
for us to make judgments about whether defiance is
a significant part of– or is complicating the
concealed carry numbers or the illegal carry numbers. Traditional crime
data, prisoner surveys, are things that empiricists
have used in this context, and I find those
unsatisfactory– at least as a
theoretical level– because my real concern is
not classic criminals who are already prohibited
in one way or another from owning or
carrying firearms, but rather the class of
people that I call refuseniks. That is, the previously
law-abiding individuals who had permits who have to
make a judgment post– or in this new, more
restrictive era– have to make a judgment
about whether they are going to carry illegally. So the cohort that
I’m interested in, or people who are sort of
familiar with the odds, who have heard the folklore
about the possibility that one might use a firearm
defensively just by brandishing and thus walking
away, those people might make a
calculation that they could continue to
carry the firearm, get some utility out of
it, and not be detected. The second category is
people living and working in rural or politically
sympathetic areas where carrying in
defiance, again, might generate a low risk
of detection or prosecution. So running through
what I think are some of the consequences–
the first possibility is that we end up with a decline
in positive defensive gun uses. And the DGU data
is contested, but I think there’s no doubt
that the aversion to a restrictive scheme would
generate a significant decline in defensive gun uses and
positive defensive gun uses, however one wanted
to categorize it. But integrating
defiance I think means that DGUs outside the home
and carry outside the home generally would
not shrink to zero. Somebody is going to cheat. The question is,
who’s going to cheat– are we happy with the result? One of the things that
I think we can posit is that these people
who I have identified as my core class of
refuseniks– who were, quote, “hyper law-abiding” is the
term that some people have used before– that is, if you
managed to get a permit to carry a concealed,
it means that you’ve gone through some vetting
that suggests that you’re trustworthy and that the
pool of concealed carriers is more law-abiding than
just the general pool of the population, which
also includes people who have criminal records. And what we find is that the
behavior of these people– I talked about this
in other pieces– that the behavior of
concealed carriers– notwithstanding the sort of
average circumstances that I’ll talk about in the second– compares to police
or off-duty police in terms of their general
propensity to abide by the law. So, question– what’s
going to happen here? Seems to me that even within
that pool of refuseniks who are, quote,
“hyper law-abiding,” you have a kind of spectrum
so that the more risk-averse individuals will
not break the law. They will fall out of the pool. But not everybody will
fall out of the pool. What happens then,
it seems to me, is that the pool of gun
carriers that I care about becomes more risky, maybe
more male, more aggressive. We can talk about whether that
makes us better off, given that we’ve had a shrinkage
of the overall number of gun carriers. The next thing that
one worries about is– and this is not so much a worry
but maybe a positive benefit– how did these people behave? Even though they
are more aggressive, even though they are
younger, should we expect that illegal carriers
will behave in the same way as carriers who had
previously been permitted? The expectation here
again is that brandishing is the preferred
option, and one should expect to see a significant
decline in those episodes where an individual
takes his concealed carry permit as authority
to behave like police. So the George
Zimmerman-type behavior, where someone pursues a kind of
police-style license and moves to a stage of enforcing
the law and attempting to apprehend someone who
was attempting to flee. Those should decline
substantially. We can fight about whether
that sort of benefit offsets the worry that we
will lose positive DGUs offsets the worry that the
general pool of carriers is more aggressive, younger,
or perhaps more volatile, more inclined to take the
risk of breaking the law. The most significant issues
in my view in this context, occur as we incorporate
defiance into this equation in the context of race. And so one thing that I posit
is that that refuseniks will be disproportionately white. Defiance is facilitated
by the Fourth Amendment, by the expectation that I
will not be randomly stopped and searched, that
I’ll have a greater expectation of the
sanctity of my home, will be respected
by law enforcement. And we know that,
like other things, that the Fourth Amendment
protection operates unequally across racial groups. We see this bolstered
in, for example, the data on Terry stops. We were talking at
lunch about this. New York, which has given
us the most significant, I think, in powerful data, about
how racially tilted Terry stops are. New York has as
dramatically decreased the number of Terry stops
since the bad old days of Michael Bloomberg. But if you look at the
proportionate or the percentage data, the proportion
of blacks and browns and whites since
Bloomberg has remained almost exactly the same. So there are fewer Terry
stops in New York now, but the same percentages that
were problematic when the New York Times was doing
its exposé 10 years ago, those percentages
are still in place. So what impact here? So the first likely impact
is that the distribution of concealed firearms,
post-restriction among blacks and browns within the
immediate boundaries of their communities,
will decline. But also, you’ll
have a sharper tilt toward those
worrisome characters– the people who are
going to be carrying in the immediate
communities are going to be those, again, folks with
a higher appetite for risk. Now, if one considers
defines to be purely a criminal
kind of activity, this is a perverse
argument, right. You’re saying,
well your complaint is that black
criminals are not going to get the same bounce,
the same positive bounce, the same benefit of the
doubt from law enforcement as white criminals. And that’s a very odd thing
to argue except that that’s the core of a lot
of the criticisms today of mass incarceration. A slightly different set of
concerns arise in the context of the sorts of defiance
that we might expect to occur in jurisdictions where
public officials– sheriffs, county commissioners,
mayors of sanctuary cities, Second Amendment
sanctuary cities– where public officials have
already stated their intention to defy a state law. So imagine a scenario
where the state moves from a permissive jurisdiction
or permissive concealed carry regime to a more
restrictive concealed carry regime after
some litigation work that shifts the commands
that come from the courts. Assume that this reversion
is surrounded by the kind of rhetoric that I’ve
already described– the sanctuary city
rhetoric, et cetera– statements like
the kind of report that we get from the
Forbes correspondent who was describing these
sort of subtle cues or under-the-radar cues
from local sheriffs that you don’t have
to worry about this, we’re not going to enforce it. So how will that
gloss on defiance be impacted as we
integrate race? Should we expect that sheriffs
and political subdivisions who have taken this stance will
apply this sort of discretion in an egalitarian
way, or should we expect that it is
going to be tainted by the sorts of preferences that
we have seen in other contexts? Just the Terry preferences,
but more than the Terry preferences. The sort of historical
kinds of worries that we see where
discretionary regimes pop up. So this intersection
is uncomfortable and I’m not sure
exactly how far to take my claim about what we should
expect to be the results. I’m not sure if the
results are things that we should say
are likely to be similar to the sort of
blatant discrimination that occurred in
the 19th century in the application
of firearms laws, or maybe the de facto
standards that we– or the de facto
discrimination that occurred in the context
of the mid-20th century, or the simple
cronyism that was one of the impulses that
fueled the “shall issue” movement in the 1980s. But what I can say
is that for people who have a concern about unequal
administration of justice, that the combination
of restrictive regime with the defiance impulse– especially where that
impulse is endorsed by political subdivisions– sets us up for a set of problems
that could generate results that are as problematic–
or problematic in the same category– as other overt
discriminatory gun laws that we have a rich history of. PHIL COOK: All right,
thanks very much. On to John Pepper. JOHN PEPPER: –figure this out. PHIL COOK: Let me
get out of your– JOHN PEPPER: Huh? You’re fine. PHIL COOK: Yeah, but
if you want to move the podium, [INAUDIBLE]. JOHN PEPPER: Well,
I feel a little bit like I’m a fish
out of water here. I haven’t seen anyone put up an
equation and very little data, and no one’s been talking
about empirical results, but thanks for
inviting me, Jake. And everyone’s just
kind of talking, you know, seemingly off the
top of their head with very eloquent words and
I’ll do my best and I won’t use a
single equation, either, which is good for an economist. This paper is
focused on evaluating the effects of
stand-your-ground laws on crime but is really intended
as an illustration as my effort to understand why
the empirical research that has evaluated guns and crime has
struggled to achieve consensus. And so I’ll try to do both in
this paper in my presentation which is to illustrate the
sensitivity of inferences to the way social scientists
approach these questions and try to get some answers
to this important question. So “explaining a violent death
is a difficult business.” This is a Jim Wilson quote. And so we know that
firearms are involved in lots of violent events–
homicides, suicides, and other violence–
but determining how many violent events
would have occurred in the absence of
firearms is difficult, and this is what
we struggle with. There are two
problems that we face. One, which is somewhat unique
to this area of research, and the other that’s common
across most social science research. The first is that the federal
government really just hasn’t invested much in
learning about these questions, and there’s a whole history
to that which we won’t get in. But we just don’t have a lot of
good data on these questions. Now, I’m not going
to address that here. I’m just going to use the
Uniform Crime Reports. Thanks to John Donohue,
you can get lots of data from every state,
from every year, on all the crimes
reported in the UCR, and that’s just what
I’m going to use. But the problem I’m
going to struggle with is the one we all hear
about– the correlation is not causation. And the way I like
to think about this is it’s counterfactuals. So let me tell you
about the question I’m going to ask today. So today I want to know what’s
the effect of stand-your-ground on the 13 states who
adopted these laws in 2006 in the period right after 2006. So right from 2008 to 2010. So what’s the effect of
adopting these laws in these 13 states on crime,
on violent crime, right after they
adopted the laws. So that seems like a
reasonable question. And a way that we typically
measure something like that is to compare the violent crime
rates when they had the laws and then ask the
counterfactual question– difference the counterfactual– of what would have
happened in these 13 states if they didn’t have the laws. Well, that’s counterfactual. There’s no way the
data can answer that. You can collect as
much data as you want and you can never at
answer the question of what would have happened
had these states not adopted the laws. So let me tell you– let me give you some data. So here’s the data. So I should tell you about
the research that we exist. So in this world where you
have these counterfactuals, the data alone can’t
answer the question. So you get lots of– you
tend to get in this area and firearms research
lots of different answers. So John and I and maybe
Phil have done work on right-to-carry laws. This is very common
in that literature, and in this literature on
stand-your-ground laws, you get John Lott who did some
work on this question about a decade ago and found that
stand-your-ground laws increase murder– reduce murder, excuse
me– by about 10%. And more recent work
by Cheng and Hoekstra found that it increases
murder by 11%. So not only are they
different, but they’re qualitatively different. Very different implications. And so, again, this is
because you’re not just using the data– you
have to apply assumption. So here’s my illustration. So what we do is extend some
work that I did recently with Chuck Manski, we’re going
to eventually apply some weaker approaches because the
conventional approaches lead to wildly different answers
that are difficult to reconcile. So we’re going to try
to reconcile the results and provide some constructive
answers to these questions that we think have
some credibility. OK. So here is what I
wanted to get to. I was rushing to
get here because I love these two-by-two tables. You know, when we
do empirical work– social scientists
do empirical work– we tend to have
much fancier models. But this two-by-two
table, in my view, kind of distills the
issue into something that hopefully everyone
can understand. So what we’re
interested in, remember, is the crime rate that
occurred in these 13 states when they had these laws. So we have the
two-by-two tables. The 13 states I’m going
to call the treated states and I’ve got another
29 states that I’ll call the untreated
states, and these are states that never adopted
stand-your-ground laws. I just wanted to focus on
the 13 states that adopted it and another group
of states that never adopted stand-your-ground laws. And so among the treated states,
in the after period after 2006, from 2008 to 2010, I know
what the crime rate is– the murder rate– it was 6.17. So that’s telling me one
of the things that I want. I want to compare
what would have happened in these
13 states if they had stand-your-ground laws. Well I know that, it’s 6.17. But what I don’t know is
what would have happened if they didn’t have the laws. I have lots of other
information in this table. I know what the crime
rate was in the 13 states before they had the laws. From the year 2000 to 2004, the
average murder rate was 6.94. And I know what was going on in
the untreated states the law– before 2006– the
untreated states had an average
murder rate of 5.16 and afterwards they
had a rate of 4.52. But that’s not what
I’m interested in. I’m interested in what
would have happened in the 13 states in
the period 2008 to 2010 if they didn’t have
stand-your-ground laws. And so none of the information
that I have can tell me that, so I need to impose assumptions. So here are three
different models that are actually applied
in this literature with a lot of other fancy
statistical stuff going on, too. But just to get to the point,
here are the three models. So that’s the table
on the left-hand side and then the three models. Remember, I want to use
the data that I have. The information I
have will somehow say something about what
the counterfactual is. So one idea is to say, well,
why don’t we use the information on the treated states
before they actually adopted stand-your-ground laws? So that’s the period of,
you know, the early 2000s. And we know that before they
adopted stand-your-ground laws the crime rate was 6.94. Now, is that a reasonable
estimate of the counterfactual? So what I’m saying is that the
crime rate in these 13 states would have been 6.94 if
they didn’t adopt the laws. Well that would be
reasonable if you think nothing else
changed that affects crime between the early
2000s and the late 2000s. Well, that’s probably not
such a reasonable assumption. I mean, I think
most people on here would think there’s other things
happening that affect crimes over time, but this is
actually– this model is used in the literature. So it’s not so crazy
to think of it. If you do that and the
counterfactual rate that you assume is 6.94, and
the treatment effect, remember, is the difference
between 6.17 and 6.94. So if you use that
kind of model without– again, you do a
lot more, you put in all kinds of
other information– but if you did
that, you’d conclude stand-your-ground laws decrease
crime by about a point, minus 0.77. OK, so you know all of you
are probably saying oh, that’s a crazy model. Well another model you might
use is contemporaneous models. Say, well I think crime’s
changing over time, so I don’t want to
use the pre-period. But what I’m going to use
are the other 21 states or 29 states that didn’t
adopt stand-your-ground laws in the same period, OK. And then you’d use this 4.52
and you’d say the counterfactual rate– what would have happened
in the 13 states had they not adopted these laws– would have been– the murder
rate would have been 4.52. Now, that would make
sense if you believe the states that
didn’t adopt the law are the same in their propensity
for murder rates as the states that did adopt the laws,
but again, that’s kind of– I don’t know, you
may believe that. If you do that, then you say
the counterfactual is 4.52, you difference that
from 6.17, and you find a very large
positive effect of stand-your-ground laws, 1.65. So it’s a very different
answer than you get if you used the
before-after analysis. Now, a third approach that
economists or social scientists have used much more
frequently as of late is the difference-in-difference
approach, and without getting into any
of the technical details, it’s just assuming that the
trends in the treated states would have been the same as the
observed trends in the control states, or the untreated states,
and you can use that idea to back out a counterfactual. When you use that idea,
you get this number, 6.29. And I’m sorry, I
do have an equation in there– there
is a subtraction and an addition sign there. But that’s how I
get that number. The 6.29, if you
subtract that from 6.17, you get something that’s
very close to zero. So here you have these
three different models which have been applied
in this literature. You know, you’d have to– with a lot of other things going
on– but these three models have all been applied
in this literature. In one case you get a relatively
large negative effect, in one case you get a
relatively large or very large positive effect, and
in the final case– this difference-in-difference
approach– you get something that’s close to zero. So that’s the problem
in a nutshell. OK, so I don’t want to
spend any time on this, but for those of you who read
and produce empirical work, you think, well,
gosh, the literature does a lot more than this. Phil and John and I have
all done this kind of work and we don’t just look
at this two-by-two table, we have lots of
other information. You look at all the states and
you’d look at time variation, and you have covariates, and you
have fancy statistical tools. But all of these tools face this
counterfactual outcomes problem and it’s arguable that
these models are imposing additional assumptions that are
even stronger than the three that I have made– in
addition to the three. That they’re all assume–
they typically all assume some version of
these three assumptions and then they add in
some other assumptions. But they also have
more information, so there’s trade offs here. But it’s not just
that adding more data resolves this problem. You know, the counterfactual is
counterfactual– the data can’t solve it. You need a model. So the problem we face is that
our assumptions, our results, our inferences, are very
sensitive to the assumptions we make and the data we use. And because of that, we– and this is a very
controversial area, obviously– and because of that,
we tend to get results that are all over the place. So, what I did originally with
my paper with Chuck Manski and now in this paper
is to say, well, let’s not throw out the idea. Let’s see if we can come
up with something that’s informative without throwing out
the basic idea of these models. So consider the
before-after model. The idea of the
before-after model is to say the 13
states today are similar to what the 13
states are yesterday, but it’s a very
strong assumption. They’re not only similar,
it’s exactly the same. Today and yesterday
are exactly the same. Where the states
that adopt the policy and don’t adopt the policy
have the exact same propensity for criminality. Those are very strong
assumptions, right. But we could back away
from those assumptions but use the same basic
idea and say, well, it’s not that today and
yesterday are exactly the same, but they’re similar
to each other. Yesterday is informative about
what would have happened today if we didn’t adopt these laws. Or the states that
didn’t adopt the laws are informative about the
states that did adopt the laws, they’re just not
exactly the same. And so we introduce this
idea of bounded variation where we introduce this
idea of similarity. So we call this s– this is another equation. Wow, I really blew it here. So, remember we’re interested
in the counterfactual. If you’re using the before-after
analysis, you’re saying– and you’re assuming
the strict model– you say that yesterday is
exactly the same as today, then you’d say the
counterfactual is 6.94. But we think things
change over time and so we allow the
counterfactual to be 6.94, plus and minus a
little bit of noise, and we call this s the
degree of similarity. Now, where do we get this? That’s a kind of big question–
where do we learn what s is? And that’s something we
talk about in the paper, but what I want you to
see here is that this just allows us to back away
from the strict equality and instead allow for
there to be some noise. So today and yesterday
we are assuming are similar to each other. They’re just not
exactly the same. And the larger the s is,
the weaker the assumption, the more plausible
your results are, but you’re going to
learn less, right. So the larger s
is, the less you’re going to learn about the
counterfactual and then about the treatment effect. So what do we do with this? There are really two
things we do with this. One is just to remain
completely agnostic and just say, here
you are, my audience. I don’t know what
you believe about s, but I’m just going to– I’ll produce the range of
results that you can get, and you pick the
s that you like, and then you can back
out what the results are. Now, that’s not very
satisfying because, you know, maybe you want me
to tell you what s is because you think I have
some knowledge of the world, but also because yeah,
you’d like some guidance and you don’t get a lot of
results that are credible but not very informative. So we can do is kind of–
just in the interest of time, I won’t spend much
time on this– but this is the kind
of sensitivity analysis you can do where these
different v’s on the graph are just displaying
how much you know. So if s is equal to zero,
you get an exact answer. Of course, the answers
aren’t the same because they’re
different models, and as s gets larger
and larger, the bounds– these bounds get
wider and wider, and you’re learning
less and less. And at some point, even
for small values of s, you don’t even know whether the
effect is positive or negative, but we can all
believe in the effect. I mean, you may not
like the answer, but at least we can
all believe in it. We know the effect lies
somewhere between negative infinity and positive
infinity, but here it’s even better than that, right. And so I’m not
trying to be too– I mean, I’m being silly
here– but you can actually– I’m allowing– and
remember what s is allowing me to do– it’s
allowing me to back away from assuming they’re
strictly equal to each other to something less than that, and
you can still learn something. Now, it may not be exactly
what we want to learn, but it’s still something. And so it’s better than
saying the effect is on one study plus 10 and then
another study is a minus 10, and there’s no way to bridge
the gap between those. Five minutes? Yeah. So I don’t want to
dwell too much on this, but I actually think
this is quite valuable. But the other thing
we do in the paper is to try to take a
stand on what s is. And I won’t get into that here– I don’t have the time. But essentially we’re going
to use the data before 2006 before any of these states
had these laws to try to get some idea of what the
magnitude of what s might be, and we choose the most
conservative value for s. Remember, s is the
sensitivity parameter, and the larger s is,
the farther we’re moving away from an exact
answer in the counterfactual. So we do that and we
get these results here. And I’m sorry, I’m
terrible at PowerPoint, so I don’t know if this
is easy to read or not. Let’s just focus on
the violent crime results just for the
purpose of this discussion. And then I’ll kind of back away
and give you the big picture results. For the violent crime, remember
what we’re doing is all three of these models but we’re
now introducing this s, this degree of similarity. And so if you look at just
the before-after model on violent crime, the
before-after model tells you that the effect
of stand-your-ground laws lies somewhere between
minus 120 and 50. So that’s a wide range. Remember, the base here–
you don’t know this, but the base is 450, so
it’s a very wide range. And so if you’re just willing to
assume the before-after model, and you’re really uncomfortable
with the standard model, but you like my value of
s, then what you learn here is not very much. But that’s just the honest
appraisal of the information we have. And just to be clear
for those of you who are used to looking
at empirical results, this is not a
confidence interval. This is the actual estimate, OK. But then if you do the
contemporaneous model, right, then the results
are strictly positive. So here we have a very, I
think, arguably weak model. I assume very little
relative to what we do in the normal
literature, and we’re still getting a positive effect. The difference-in-difference
model is between 2 and 52, so that’s positive. And if you combine all three
models, which I can do here, I get as a result
that’s between 3 and 50. So what do we interpret then? As you say, the effect of
stand-your-ground laws– I don’t know the exact answer– but I can tell you
it’s a positive effect. It lies somewhere
between 3 and 50, which is roughly an increase
of about one to 15%. I mean, so there’s a
big variation, but still a positive number. When you get to a murder, you
get a point estimate– a 0.2– which says that the effect
of stand-your-ground law is to increase the murder rate
by 0.2, which is a 3% increase. So it’s a little
less– that’s less than what Cheng
and Hoekstra found but still consistent with their
finding of a positive effect. And then if you look
at robbery and assault, the results, even when I
combine all three models, are uncertain. We don’t know whether the effect
is negative or positive, which is again somewhat consistent
with what Chang and Hoekstra find in their paper. So I’ll just wrap up now. I don’t know if I have
any time, but I’ll just– so we really– I wanted
to approach this problem by highlighting, for you,
the kind of trade-offs between assumptions
and credibility and also the sensitivity of
the results to the assumptions you make, and to explain why
we get such different results, or to illustrate that. And then apply these new
weaker-bounded variation assumptions to get what I think
are credible and informative results. PHIL COOK: Terrific. Thank you. So the third paper–
we’ll go right down– is John Donohue of Stanford. And we got to– SPEAKER 1: Go to the next one JOHN PEPPER: I’ll just go
to the next, do you think? [INAUDIBLE] JOHN DONOHUE: I’ve been a
little sleep deprived this week, but as soon as John
put up those equations it was like adrenaline going
into an empirical economist’s veins. JOHN PEPPER: I
could’ve put up more. JOHN DONOHUE: So I
picked up quite a bit, thank you for that. So this is a little bit of
an unusual paper for me. It really has sort
of three dimensions. One is sort of a cultural
commentary point, one is more of the empirical
evaluation of the type that John was just
going through, and then the third
has a legal part. But I do think the swerve is
sort of an interesting metaphor here. As you’ll see– and a number of
people have sort of commented on this– it’s hard to think of the gun
problem in some uniform way because we actually have
many different gun problems– everything from
suicides and homicides and accidents and street
crime and mass shootings. And they’re all different– sort of require
different approaches if you are going to
try to solve them– but with 300 million guns,
everything becomes difficult. And as John’s remarks
really highlighted, it’s not easy to assess what
the costs and benefits of gun regulation are, and
of course the politics are very challenging, and
now we have the overlay. So we’re really at a
very interesting moment in American history
where the public has now been very energized– largely by the mass shootings–
to want more gun control, just at the moment when
the Supreme Court now has four solid votes for sort
of ending almost all gun control and the fate of the
Second Amendment will pretty much lie in
John Roberts’s hands, as long as Ruth Bader
Ginsburg survives. So here’s the Second Amendment. Interesting. I like the way the NRA thinks
of the Second Amendment. They put the pistol over
the first half of the Second Amendment so you don’t see that
“well-regulated militia” part. And this was a
pre-Heller document and then Scalia said, yeah I
sort of like the NRA approach, so we’ll get the first
half of the amendment. But if you look back over
time, the NRA actually had a very different view. This is the NRA president
testifying for the Federal Gun Control Act of 1934 and
he said, “I’ve never believed in the general
practice of carrying guns.” The promiscuous
toting of guns should be sharply restricted
and only under licenses.” So imagine an NRA
president saying anything like that today– that would
be some communist that we could barely tolerate. “There’s no reason
why on the street today a citizen should be
carrying loaded weapons.” Guns are a “ridiculous
way to solve problems that have to be solved
among people of good will.” I mean, you could
just hear the words of AOC being attacked as a
communist for saying this. That was Ronald
Reagan in May of 1967. “The Second Amendment has
been the subject of one of the greatest
pieces of fraud– I repeat the word fraud
in the American public by special interest groups.” Referring to their
constitutional assessments. Again, this is in a
pre-Heller period. That was the very conservative
Republican Chief Justice Warren Burger speaking in 1991. Another conservative Republican
Erwin Griswold sort of mimics the same concept. So you see there’s
been an enormous change in the attitudes of
the Republican Party on these questions. Today, here’s a more traditional
Republican legislator who has introduced a bill
saying that every resident of the state of Missouri shall
be compelled to buy an AR-15 and also should have
a legal compulsion to buy a small arm that
he can carry with him. So you get the sense there’s
been a change over time in the way people
think about things. This is from a decision
in May in which a federal judge in California– which was alluded to earlier
today in Bob Spitzer’s presentation– struck down California’s
restrictions on high-capacity magazines. And he said, “This is
not to be lamented. It ought to provide reassurance. It is the proudest boast
of our Second Amendment, jurisprudence that we
protect the citizens’ rights to keep and bear arms that
are dangerous and formidable.” So the shift in time has really
been quite extraordinary. But let’s look at
some data here. My source for this was Louis
Klarevas’s book Rampage Nation. And he had put
together something that had suggested the Federal
Assault Weapon Ban had actually diminished mass shootings and
deaths from mass shootings, and then it had risen
sharply thereafter. So, you know, for all the
reasons John mentioned, you’re always skeptical
about other people’s data. So we got a different data set,
tried to do our own analysis, and lo and behold, the
Klarevas calculus did hold up, and indeed we
extended the analysis for five additional years. And what you do see is that
for the 10 years of the Federal Assault Weapon Ban, there was in
fact a fairly substantial drop in both the number of
occurrences as well as the deaths and the very
sharp rise afterwards. There were some
early evaluations of the Federal Assault
Weapon Ban done in 2004 that said we really
can’t say too much about what the Federal Assault Weapon Ban
did because crime was trending down. And if you look at
this dotted line here, this represents the
violent crime rate. And you can see over this
whole period for the nation as a whole, violent
crime was trending down, and yet for the mass shootings
and the number of deaths from mass shootings, you
see the very profound drop during the 10 years of
the Assault Weapon Ban, and then the sharp
rise thereafter. And in this final period– which is only five years long– we’ve already surpassed
the previous decade in terms of the
number of deaths. And the projected level,
if this rate continued over a full decade,
is we’d be way up here five years from now. So it’s sort of a bad
trend for mass shootings even though violent
crime is trending down. So you get a sense
of the way in which certain elements
of the gun problem are getting better in
terms of violent crime overall, but for mass
shootings clearly moving in the wrong direction. OK. So it’s worth noting that every
gun massacre in the last five years that makes it
into this calculus– where six or more people died– used weaponry that was banned by
the Federal Assault Weapon Ban. So that means they either
used an assault weapon that was specifically banned or high
capacity magazines, or both. OK. This is just sort of showing
that the number of deaths per episode dropped during
the Assault Weapon Ban period and then rose
sharply thereafter. So you can see that we had
quite a substantial jump, you know, basically
doubling of the number of deaths per episode since
the Federal Assault Weapons Ban ended. This just sort of
highlights the fact that you could not
have a mass shooting problem at the time the
Second Amendment was written. It would take 30 seconds
to load the top gun. The bottom gun– obviously
can kill a lot of people very quickly as we saw in Dayton
where the shooter was brought down within 30 seconds and
yet the number of people shot was quite considerable. So we wrote up some of these
results in the New York Times. September 4th, the NRA put out
this headline just this week, “Fake News Alert.” And so as everything is in
the Second Amendment arena, there’s a lot of controversy. But the NRA stated, “For some
reason, Donohue and co-author Boulouta chose to use six
fatalities of the minimum instead of the common
definition of four or even the Obama era attempt to
reduce the threshold to three. Reason Magazine has a pretty
good guess why they did it– it obviously was
cherry-picked to get results that they liked.” But let’s look what
happens if we look at the four-person standard. Boy, the graph looks a
little similar, doesn’t it? So no, no, no. There was no cherry-picking. I used six because that’s
what Louis Klarevas had used. And there were
obviously arguments that one might be better,
but essentially the higher the standard you use– it is harder to get
to a higher standard if there is an assault
weapon ban in place– so it does make
it more dramatic. This is one of the
problems with what has happened in the aftermath of
the Federal Assault Weapon Ban. In 2005, there was a
federal immunity statute that essentially
insulated the gun industry from this type of advertisement,
which they probably would have felt a little fearful
of with tort lawyers coming after them. This was the gun
used by Adam Lanza and his [INAUDIBLE]
card had been reissued. This is, again, another
NRA item again trying to convey this idea– you know, guns will
not only keep you safe, but promote your
overall good feelings. The 19-year-old Parkland shooter
was very moved by that image and shared it on Instagram. And his language in
his statement I thought was meaningful. He said, “With the power of the
AR, you will know who I am.” So clearly guns were
influencing his notions and he said he had been tired
of being called an idiot. Somebody wrote to me in the
aftermath of the New York Times piece and said you know you
don’t know what you’re talking about, the top gun was
banned and the bottom one isn’t, so the Federal
Assault Weapon Ban could have played no role. This gun here, though– which is legal in the states
that banned assault weapons– has never been used in
any of the mass shootings. And part of the reason why
I conjecture is this kind of– it’s called
the Mini Ruger– it’s sort of a weenie-looking
gun compared to that gun. And if you’re Adam
Lanza and you really want to be the tough
guy on the block, you’re much more likely to
feel your sense of grievances will be addressed
if you’re holding that top gun than if you’re
holding this bottom gun. So, I do– whenever I
hear anyone say the word law-abiding citizen,
I usually think they’re on the wrong track. It’s usually use like you
should never take guns away from law-abiding citizens,
because all of these guys were law-abiding citizens until
they started the mass shooting. And at a meeting of the
National Institute of Justice, the new head David Muhlhausen
was former Heritage Foundation employee, mentioned
that he didn’t want to support any
research that didn’t involve randomized experiments. And it’s sort of a
noble aspiration, but if you’re talking
about gun issues there aren’t a lot of
randomized experiments that are easy to do. So I mentioned, you
know, that means you’re not going to
support any gun research. And he said, well I
would never support any research that
would take away guns from law-abiding citizens. Again, I think that’s
the wrong track. Here is some
language that I think might be useful to think about. “The vigor of government is
essential to the security of liberty. The contemplation of a sound
and well-informed judgment, their interest can
never be separated. That dangerous
ambition more often lurks behind the
specious mask of zeal for the rights of the people.” This is not a
sentiment that most of the originalists of
the NRA crowd likes, but that was from the
first Federalist Paper. I think it probably
captures a little bit more of the way the
founders of the country would have thought about
not only the Constitution, but general [INAUDIBLE]
2nd Amendment. So we now have seen not only
this shift in the rhetoric, now I’ll just give
you a sense of a shift in the actual legal
situation across the country with right-to-carry
laws– laws that allow you to carry concealed
handguns if you get a permit. And so this was as late as 1986. Red areas all did not allow
issuing right-to-carry permits at that time, and then
jump ahead to 2019, and now virtually every
place is shall issue– the sort of bluish, purplish– and the bright green
requires no permit at all. Anybody can carry there. So I was just on a
trial in Missouri where now anyone
who’s 19 or over can just carry without a permit,
and the University of Missouri has tried to say we’d
rather keep 19-year-olds off the University of
Missouri campus in conformity with our 65-year-old
ban on guns on campus. And so that has been
challenged by the NRA and the attorney
general actually backed the NRA position
against the university. So we’re still awaiting
the results of that trial. So not only is there a lot
of litigation on this issue, the move to state permitless
carry seems to be growing. And even things like the federal
legislation on reciprocity– so that if you have
a permit in Arkansas, you can go to the
Boston Marathon and carry your weapon if
that law were to take effect. And of course, the
Supreme Court decision might say you have a
Constitutional right, so they would make
permitless carry perhaps the law of the land. But, again, what can go wrong. Pretty much this is like
the worst-case scenario for right-to-carry. Two guys were having
a tailgating dispute, they pulled over to a nearby
car wash, stepped out, they were both carrying
under right-to-carry permits, and they both died in
the hospital that day. Sometimes John
Lott say, oh there are very few permit revocations. Neither of these guys
had their permits revoked because there was
no need for the revocation in that case. And we heard earlier
about the Philando Castile case, which I think is an
important case to remember. You can actually
see this on film. Very polite commentary
between the police officer and the driver until he
mentioned by the way, officer, I’m carrying with a
right-to-carry permit, and shortly thereafter
he was shot seven times. He was reaching for the
wallet that the officer asked him to get his license, and that
tends to be a dangerous thing. And in part it’s because
the threat to police comes from only two
possible sources. One, they get into a
car accident, or two, they get shot by someone. Police virtually never die from
anything other than those two things while they’re on the job. But this, I think,
gives you a little bit of a sense that because of
the gun culture in the US, the nature of policing is vastly
different than it is elsewhere in the world. So in the first 24
days of 2015 police fatally shot 59 individuals,
which was more than the number shot in England and Wales
over the past 24 years. So we’re often shooting
people and killing them by the police at 100 times the
rate of our comparable nations. Just to give one sort of graph
on one of the types of analysis that John was
referring to that you can do with the right-to-carry
laws– this is a data analysis. And the nice thing about
this particular model that I sort of like
is the fact that this shows you what is
our model predicting the effect that the
law is in the 10 years prior to adoption of
a right-to-carry law. And you would
expect it would have to be 0 because the law wasn’t
in effect prior to adoption, and lo and behold, it
looks pretty close to zero, and then right at the time
of adoption starts rising up. So if you believe this
particular panel data model, it suggests that over the
course of a 10-year period you’re getting close
to a 15% increase in violent crime associated
with a right-to-carry laws. You can also look at the
same question using a newer technique that actually
wasn’t even in concept I think at the time that the
National Research Council report on right-to-carry
laws came out in 2004. This is a newer technique from
some economists at Harvard that gives another way to sort
of address John’s concern, because as he mentioned, when
a law passes we don’t know what the counterfactual is,
and the synthetic control becomes a technique for trying
to identify what that is. Both of these approaches
have problems and limitations with them. The interesting thing for
me is that they both led to the same conclusion. And so here you can see the
synthetic control fits rather well in the
pre-adoption period– the blue represents what
the synthetic control is, and it tightly fits over
this pre-adoption period– then Pennsylvania adopts
a right-to-carry law, and that extends
it to Philadelphia. You see crime spikes and
the synthetic control would have predicted a drop. Same thing for Texas where
the synthetic control predicts that crime would have
taken a better path had they not adopted
right-to-carry. This is another
illustration with Missouri. The top line shows
what actually happened and the blue line shows
what the synthetic control prediction would have been. So again, just amalgamating
over the 33 states for which we have
estimates comes out very similar to the
panel data estimate– maybe an increase of 14% or so
over 10 years after adoption. Interestingly, I was able to
do that same sort of panel data and synthetic control analysis
looking at the question of what is the percentage of robberies
committed in these states after right-to-carry adoption,
and the panel data estimate suggested an 18% increase in
the percentage of robberies committed with guns. And the panel data
estimated was even higher, yet right-to-carry
laws were not reducing the number of robberies. So I think it makes one
point that sometimes gets lost in the debate– the NRA position is, you know,
how could it be bad to allow– how could it to be bad to
allow law-abiding citizens to carry guns? At the very least, it’s
a very static analysis to think that way because the
criminals respond to more gun carrying and this
happens to lead to results that are adverse. Also in the mass
shooting arena as we just saw, as the gun
carrying is increased, we haven’t seen right-to-carry
laws achieve the goal that was articulated in
the early stage– if everybody’s
carrying guns, they’ll take out the mass
shooters so you don’t have to worry about
that problem anymore. We’ve seen that’s not
true and it turns out even for police it’s
a dangerous enterprise to try to take out these
mass shooters, which is one reason why armed citizens
that aren’t military or police play virtually no role in
dealing with the mass shooting problem. This is sort of the
imagery, though, that– the smiling face
of the guy who’s protecting us all and we
should feel good that he’s carrying around his gun– but I think this is
the way in which we need to alert the public
about the nature of risk. Because Nancy
Lanza living in one of the safest cities
of North America– Newtown, Connecticut–
felt she needed to have an arsenal to
protect her and her family from the hordes that
might come to attack her. And she believed
that very powerfully even though she
lived with Adam who hadn’t left his room in a
year and was burning himself with cigarettes, that she made
her weaponry available to him. And many people, including Nancy
Lanza, paid for that folly. Almost the identical
pattern we’re seeing for the mass killer in
the unquote community college scenario, where he
killed lots of people at this Oregon
University in 2015. He was struggling
with mental illness that his mother
was fully aware of. He’d even threatened
her with a shotgun, yet she boasted that the
arsenal that they had would keep them safe and
sneered at gun safety efforts and had written, “I keep two
full mags in my Glock case, and the ARs and AKs
all have loaded mags. No one will be dropping
by my house uninvited.” Except the police came by to
look into why her son had just killed lots of people. This is Marion
hammer of the NRA, a powerful lobbyist in Florida,
proposing or arguing against the NRA-proposed– or the Florida-proposed
Assault Weapon Ban– and she said, “how do you tell
a 10-year-old girl about a Ruger 10/22 with a pink
stock for her birthday, that a rifle is
an assault weapon and she has to turn it
over to the government or be arrested for
felony possession.” So I do think you can see the
nature of the public attitudes about guns has changed
in a profound way. The sense that I
always need my gun and it always
needs to be loaded, is what I need to
protect me, has changed in a very profound way. And the sense that there are
dangers of guns everywhere has not been nearly as
profound as it should be. This is a case of a
girl just a short while ago who was killed
when her mother was about to take her and
her four-year-old brother to a Little League
game, and the mother had run back into the
house to get something, and the mother kept a loaded
gun in the gun console, which the four-year-old boy found and
shot his sister in the head. And the police immediately
said no charges will be filed and we know that the Lord’s
loving embrace of the girl will give the family some peace. But I do think this is actually
a bigger concern for most people thinking about
buying guns than the thought that they’re going to
be the hero stopping some mass shooting episode. PHIL COOK: Great. Well, thank you very much
to all three speakers, so– [APPLAUSE] Let me say that we do have
a few minutes if somebody has a question, please
raise it to any one of the three or all three. As soon as the grinding stops. Yes, Darrel? AUDIENCE: [INAUDIBLE]
what’s the difference in your model [INAUDIBLE]
between violent crime and the subsets of
what [INAUDIBLE] further assaults [INAUDIBLE]. JOHN PEPPER: The UCR just– they have– that’s
the sum of those, plus some extras including rape. AUDIENCE: OK. [INAUDIBLE] And the other is
something [INAUDIBLE] you’re educating the media as a
consumer of this kind of work. [INAUDIBLE] I think you had
said in your paper that the kind of statistical
methodology and trying to create this kind of
[INAUDIBLE] plus or minus s is somewhat new in
terms of the modeling. Can you just speak a little bit
about is this something that– that other people use
in other kinds of areas. Is this something that
specific with this kind of econometric research
or how sort of common is this [INAUDIBLE] statistical
tool now [INAUDIBLE].. JOHN PEPPER: Well, so these what
we call partial identification methods have been
developed for a long time, but seriously in
about the late 1980s. But this particular tool
is only the first time introduced a year ago. Now the– I mean– so there’s a lot of theory
the applications of this are not that wide because
for a couple of reasons. One is that you
don’t get an answer, and I think most people
like to have answers. And the other is that it’s not
integrated into many software packages, so it’s hard. You have to kind of
program stuff in. But it’s very commonly
used in a part of economics called industrial organization. It’s less common
in areas like this and I’m one of a
handful of people that are doing this kind of stuff. PHIL COOK: Can I ask
Nick a question here? So I I’m interested in this word
defiance, which certainly grabs your attention when you use it. And I understand that
with sheriffs or even with legislatures who are
defying the federal government. But I had a much
different impression about ordinarily
individuals who are suddenly told by the state of Colorado
that they now have to– if they want to sell a gun
to their next door neighbor they’re going to have
to go to a gun dealer and pay that person $35. It’s not defiance to ignore
that law any more than it would be to go 15 miles over the
speed limit or you know, it’s just convenient
to ignore the law if you can get away with it. So you’re obviously conveying
another mechanism that leads to widespread ignoring
or violation of the law, and I’m curious about that. NICK JOHNSON: So we can call
it lawbreaking, and indeed, in the paper I acknowledge that
you may just say that whenever prohibition day comes–
whatever the prohibition, whatever the restriction is– that anyone who fails to comply
is just a common criminal. In the context though
of supply controls where I think this is a
more significant disruptor– I’m working on
another paper that talks about this phenomenon–
that’s what’s happening. Question, what’s
happening in New Jersey, in New York, in
these places where we see that there are very,
very low rates of compliance? Well, it sets up whatever you
call it– defiance or just illegality– it sets up a
subsequent set of problems. So what about the
subsequent holders is my next question– that
is, the people who are heirs to all of this contraband. There are a variety of aspects
of our mens rea theory– so cases like
Staples and a couple of the child porn cases– that suggest to us that
people who find themselves subsequently in possession
of these items of contraband that have been withheld
by the refuseniks as an act of defiance
or simple illegality, that they are in a
spot that is going to be very hard to prosecute,
frankly, because how do you tell– visually– whether
you’re in possession of a 20-round magazine
or 30-round magazine or one that has been modified? So the broader point that I
have regardless of the nomencl– regardless of the label– is that the phenomenon of
defiance or noncompliance generates a series of
enforcement questions and policy questions that I
think we make a mistake if we ignore them, because we end
up with results that are not what we hope for just
in terms of design. And the statute says,
thou shall not and then presuming that
everything after that is a story of happy compliance. PHIL COOK: Yeah, yeah, no. Good intentions are not
enough, unfortunately. Thank you. We have a 10 minute
break, I guess. So here we go. [APPLAUSE]

Leave a Reply

Your email address will not be published. Required fields are marked *