The Perfect Crime: A New Freakonomics Radio Podcast

Listen now:
(Photo: Quinn Dombrowski)

(Photo: Quinn Dombrowski)

This week’s podcast is called “The Perfect Crime”: in it, Stephen Dubner describes a way to kill someone without any punishment. (You can subscribe to the podcast at iTunes, get the RSS feed, or listen via the media player above. You can also read the transcript, which includes credits for the music you’ll hear in the episode.) But let’s be clear: Dubner isn’t suggesting that anyone actually try this. In fact, the problem is that too many people are doing it already.

So what’s “the perfect crime”? It turns out that if you are driving your car and run over a pedestrian, there’s a good chance — especially if you live in New York — that you’ll barely be punished. Why?

We hear from Lisa Smith, a former prosecutor and now a law professor, who tells us that just 5 percent of the New York drivers who are involved in a fatal crash with a pedestrian are arrested. As it happens, New York has particularly narrow standards for conviction in such cases; there is a lot of variance among states.

Throughout the U.S., there are more than 4,500 pedestrian deaths a year, about 14 percent of traffic fatalities:


In New York City, meanwhile, pedestrian deaths make up an astonishing 52 percent of traffic fatalities.

Who’s to blame for all this pedestrian death? Let’s take a look at data from the National Highway Traffic Safety Administration’s Fatality Reporting System. While 25 percent of pedestrian deaths are attributed to the driver’s failure to yield, pedestrian behavior looks pretty bad too — they lie down in the street, dart into the road, cross improperly, and so on:


But these data may not reflect the entire truth. Dubner speaks with Charlie Zegeer, associate director of the Highway Research Center at the University of North Carolina, who says it’s hard to accurately determine the cause of many pedestrian deaths:

ZEGEER: The reason we don’t know is because the information we have to make that determination is essentially on police crash reports. And so oftentime the only witness is the surviving driver. And so the police officer only hears that one side of the story.

This leads to a conversation about a core issue: who do the roads belong to anyway? You’ll hear from Robert Noland, director of the Voorhees Transportation Center at Rutgers; the transportation historian Peter Norton, author of Fighting Traffic: The Dawn of the Motor Age in the American City; the Harvard “urban economist” Ed Glaeser; and a couple of New York trauma doctors, Spiros Frangos and Stephen Wall.

The doctors tell us that about 25 percent of the trauma patients who come through Bellevue are pedestrians who’ve been struck by cars. Frangos and Wall have written a series of papers on the topic, including “Vulnerable Roadway Users Struck by Motor Vehicles at the Center of the Safest, Largest U.S. City.” So what can be done to keep New York pedestrians from dying? Frangos and Wall propose — only half-jokingly — helmets for everyone:

WALL: From our data, I think all pedestrians should be wearing helmets. But who would really want to wear a bike helmet when they’re walking, when they’re going out for a date, and that’s also the reason why they don’t wear them when they’re riding their bicycle. But there’s some truth to it.

Eric M. Jones

So why not just compare the number of people who DESERVE to be run over by state? I can think of a few.


One thing to consider on the increase in pedestrian fatalities is this dramatic increase in the safety design of cars. If you look at the graph 2003 and 2012 end with ~4,700 fatalities. While the total fatalities has decreased. This is possible correlative evidence that with the advent of releasing engine blocks, side panel airbags, rotating wheel drive for slick surfaces, and the auto brake sensors have exponentially increased the value of life to the driver of the vehicle. Also, because of the steady rise in gas. The top heavy gas guzzlers have(had for awhile) been abandoned for the fiscal friendly sedan, which is half as likely to flip. While no new mass market technology has increased the value of life for pedestrian victims.

I think the best way to reduce pedestrian would be something akin to the incorporation of the bullet train. Provide commuters on the outside of Manhattan a cultural hub that would regulate transportation in a manner as fast/faster than vehicle at the same cost.



Because in the age of the affordable, safe, comfortable, and stylish car, it's so easy and inexpensive to implement mass transit options that people will actually use!


Never claimed it would be easy(or likely) but the fuel costs of bullet trains would keep costs down on the commuter to as much or less than the current value. It would be mildly expensive to initially implement but the preference of cars is that they are a) convenient(which is why I recommended a centralized hub outside of Manhattan) and b) faster that current models of mass transit(buses, current train, etc.)


Seems to me that pedestrian fatalities are making up a larger share of total traffic fatalities because cars are getting safer, and walking isn't. If you look at the data, the number of annual pedestrian fatalities hovers around 4700, from 2003-2012 while the total number of traffic fatalities decreases from 42000-33000. That tells me that for some reason there are 10k less non-pedestrian fatalities, so driving has become safer (probably due to improved safety measures on cars), while walking has remained at about the same level of safety. This may back up your point that pedestrians should start wearing safety gear if we really want to bring that number down.


Driving has become safer... for car occupants. It hasn't become any safer for cyclists and pedestrians hit by drivers, the equation of kinetic energy is still the same, one half of mass times speed squared. Technology can't change that (though there are ways to mitigate it, like a pedestrian crumple zone to reduce acceleration during collisions).

Technology has made driving more forgiving for car occupants by having their car protect them in case of a crash. But it hasn't made car drivers drive more carefully, only a culture change through policies prioritizing pedestrians and cyclists could achieve that.


Driving has actually become safer for pedestrians as well through things like back up cameras better breaking, better headlights and systems that will automatically stop the car before it hits something. In fact, I read recently that in the next few years, all cars will have to provide a means for the driver to see the area immediately behind the rear bumper to keep drivers from backing over children playing behind the car.

However, it is likely that cars have gotten safer for drivers more quickly than they have for pedestrians. But this is what makes the primary statistics relied upon in this episode so misleading. The statistic of pedestrian deaths as a proportion of total traffic deaths is very dependent on the number of overall traffic deaths.

Presumably, if the number of pedestrian deaths are going down, then life is getting safer for pedestrians, but if the number of occupant deaths is going down faster, this ratio will go up, which according to the logic of this episode would incorrectly mean that things are getting more dangerous for pedestrians.

By the way, this is pretty much what has been going on in NYC and why the ratio of pedestrian deaths to overall traffic deaths is so high in NY.



This totally applies in LA considering the amount of hit&run incidents in LA (which is an epidemic, especially amongst cyclists).


You did not cover automobile designs to make them safer for pedestrians. These could be through redesign of bumpers, hoods and through crash avoidance like automatic braking when the vehicle detects imminent impacts w pedestrians (using IR or vision cameras). Europeans have taken the lead on such initiatives as well.


This has already happened. European Union pedestrian safety standards have made their way to most cars sold in the US. Cars in recent years have higher bumpers, higher and softer hoods, higher cowl heights. I don't approve.


Why are bicyclists asked if they were wearing helmets when they were hit by a car?
How does that help us understand the causes of crashes?


Very simple. The victims are blamed and this has been the case since automobiles "own the road."


Automobiles certainly do not "own the road".

Pedestrians have right of way on driveways, zebra crossings and walkways at controlled intersections. Legally, you mow them down at your peril.


Gah! This episode left me wanting to shout at my radio. There were three major flaws in this episode:

1) Looking at the ratio of pedestrian fatalities to all fatalities in New York has a major denominator problem. I expect that New York City has very few occupant fatalities by most reasonable measurements because there are very few places in New York City where cars can drive fast enough to have a fatal accident. (I was hoping to find a comparison of occupant deaths per mile, but could not do so). If you have very few occupant deaths, then pedestrian deaths are going to seem huge, even if New York City is relatively safe for pedestrians.

2) Killing someone by shooting a gun into the air is not comparable to killing someone in a car accident. When you shoot a gun into the air, even if you are not intentionally killing the person, you are still intentionally doing something tremendously stupid On the other hand, I suspect most traffic violations are not intentional, but rather a momentarily lapse of attention. The comparison does work with driving drunk, since there, you are choosing to do something stupid, but that is illegal.

3) It is unlikely that harsh criminal punishments would decrease the number of traffic fatalities. People don't fail to see a child crossing the street because they don't have sufficient incentives, they do so because they had a momentary lack of attention. Also, there are already substantial incentives for not running over a small child, including the wrongful death lawsuit you can expect for doing so and the lifetime of guilt, not to mention that the same sort of inattention that causes you to run over a small child also causes you to get into an accident that could kill you as well.

4) Saying that walking drunk causes accidents is not the same thing as saying people who walk drunk deserve to die. It is just saying that the thing that caused their death was their own negligence, not the negligence of the driver.




The biggest improvement we could make to the safety of everyone in cities is to get rid of the crazy notion that the best form of personal transportation is a 2-ton farting metal box on wheels. Failing that, a serious safety improvement that would lower driver inattention substantially would be to ban seat belts and replace driver air bags with a driver flick-knife.


Your suggestion really isn't productive. There are many circumstances where mass transit is a better form of transportation in the city than cars, but that is far from always the case. Among other things, trains stop running at some point and run only once per hour at other points. The truth is that, at least in NYC, there is a very low chance of a pedestrian being killed in a traffic accident, and a substantial portion of the time, it is a result of the pedestrian doing something stupid.


I can't seem to download nor stream this episode. I'm technically literate and have tried three devices and three programs, including attempts to stream/download from the link on your site. I'm omw to take a tough biology final and really wanted it for my commute. I'll muddle through somehow ;-) and I'm sure it will get figured out! Love you guys! Thanks for what you do!


They don't make it easy to download, do they? They want to steer traffic to iTunes where there's money to be made. For anti-dark side (that is to say anti-Apple) people who abhor iTunes: What works for me (in Firefox) is click on the RSS Feed link, right click on the Play Now link, and click Save Link As. The .mp3 is then downloaded.


I'm not exactly sure how one "requires attention" when it comes to driving. Clearly a lot of the cause is negligence. It's humorous to me that in the wake of that NY commuter train derailment, the engineer of the train was immediately drug tested, phone records subpoenaed, and their license revoked baring review. Why is that also not the standard for driving related deaths?

I'm going to stick to planes and trains for travel because at least if I die on one of those, my death would receive many more times the attention and scrutiny than if I were just bulled over by some SUV walking to get a sandwich.


Also, the other sad truth ignored by the "freedom machine" lovers and dependents is that people who drive less or don't drive end up subsidizing the roads for those drivers on roads paid for by general revenue sales taxes. After all the regressive deductions and subsidies for the extra infrastructure required are summed up, it's amazing how the suburbs are the biggest moochers of tax dollars of anybody.


I drive in NYC and I am astonished just how many people walk directly in front of my moving car, with the green light. The impatience of pedestrians, the use of cell phones while walking and the general expectation that cars will stop for you are contributing to these high numbers. Drive for one day NYC and count these instances...then it will become clear how these numbers skew so high.


Agree. I walk in NY, and I cannot believe the impunity with which pedestrians stroll out against the light on every corner. Or the resignation with which the drivers accept the situation.


so... the take away from data that clearly shows just how dangerous cars are is that everyone should just wear helmets? of course it's unfathomable to recommend that maybe we should be making it easier for the vast majority of people to get around by other means and that perhaps ROAD DESIGN could be a factor in vulnerable road user safety - OR - maybe we do like northern european countries and require much stricter standards for obtaining drivers licenses...


I wonder if in the new book, the common thread of cost/benefit analysis comes up. And as maddening as it can seem that people are getting ran over with no consequence, it could be the sad truth that the actual "tax" on negligence isn't really worth all that much and our choices are to accept that fact or not to engage in the activity trying to reduce the risk to nearly zero.

I made the point to my friend that even if you value the lives ended annually by guns at $5 million dollars, that only works out to about $300 per gun (assuming about 20,000 deaths and 300 million guns) in externality cost, I'm sure gun owners would hugely agree the benefit is greater than that. Levitt made a similar point in the Marijuana vs Alcohol episode about alcohol having a greater societal benefits than its cost, and I'm left to wonder if that wasn't the very same cold cost/benefit analysis that angered so many in the abortion/crime rate decrease parallel drawn by Levitt as well. So, I'd bet a similar value of death per car is true for the ~30,000 traffic fatalities which makes it completely rational why it's the perfect crime, or a crime of around $500 is attributable costs.

The truth is a bummer summer, and to quote Slim Pickens in Blazing Saddles "I am depressed."



One thing I did not notice in this was a breakdown of pedestrian deaths into those hit by cars while in the roadway, and those hit on sidewalks. I would bet the second number is considerably lower than the first.

As for the deaths related to e.g. texting while walking, perhaps we need a new category, 'unintentional suicide'?


Ugh!!! Well intended but utterly misses so many points. A myopic picture of such a vast problem. And incredibly disappointing that the voices of pedestrian and bicycle transportation advocates were not included at all in this program.

The bike/ped advocacy world is probably the best clearing house of this data and analysis. Transportation Alternatives ( could have dropped a tome of meaningful data analysis on these guys. But instead they just bounce back and forth between wonks, working in silos with no connection to each other.

It's great that the ER is collecting data from pedestrian victims (although the helmet question is once again a red herring if your trying to understand the CAUSE of crashes) but it's flawed in the same way that the data collected from only the driver is if it's not cross referenced.

As a result almost no mention was made of probably the #1 mitigating factor in all crashes, but certainly in the frequency and severity of bike ped crashes: SPEED. Also absent was any meaningful mention, (if at all) of best and easiest ways to address the issue of speed:

1. Redesigning roadways to reduce vehicle speeds and prioritize pedestrian safety
2. Automated enforcement of speed and signal compliance through speed and red light cameras

Instead we end with an E/R Dr., someone not at all qualified to speak on transportation policy, suggesting what this podcast loves: a wacky left field solution, that we put helmets on pedestrians, because yeah, that's the problem.

Gunfire a problem in your neighborhood? Don't look at the causes of poverty, crime and easy access to firearms, no, just tell everyone to put on kevlar and accept things as they are. Way to blame the victims guys...



Helmets? "Victim blaming is a subtle process, cloaked in kindness and concern."
-- William Ryan


This podcast resonates with me, because here in the Bronx I cross the street at my own peril. Am not particularly fast on my feet or all that coordinated, so my huge fear is getting killing by a negligent driver.

Drivers in NYC, do NOT yield to pedestrians as a rule. And I can be crossing the street comfortably, having seen no approaching cars, and then all of a sudden I'm being intimidated by a car going over 30 miles an hour in what probably should be a 20 mph residential neighborhood with a lot of children.

Always cross with the light, but that doesn't protect anyone. This conversation is basically about two things: Speeding and distracted driving (texting, cell phone calls, eating, checking a facebook page, etc.)

Steve Magas

If you ask someone "Who do think is killed more often, people who ride bikes in traffic, or people who fall down" you'll generally hear that those pesky bike riders MUST die in alarming numbers. Except they don't… cyclists are 1-2% of all traffic fatalities. Cycling IS safer… sort of… If you break down the numbers by decade, each decade shows fewer cycling fatalities than the one before. However, in 1975, the "worst" year for cycling fatalities, 2/3 were KIDS. Today's kids don't ride - and adults are riding more than ever - so the total fatalities are "down" from 1975 … significantly… but ADULT cycling fatalities are way UP from that 70's era...

HOwever, those who fall down die in alarming numbers - in Ohio we average around 16 cycling fatalities - and 900+ fall down deaths each year - nationally there were 26,000 fall down deaths [unintentional] in a 2010 CDC report & 600-700 per year the past few years. - particularly at risk are those over the age of 65 - so, my response to those who want bike helmet laws is to suggest we start helmet-ing the elderly since they die falling down 40+x more often than folks who ride in traffic! When you turn 65, you are assigned a state-approved helmet…



You were wondering why all these pedestrian deaths are not causing more outrage. I think it's the economics of it; ie: the income of those killed vs those doing the killing. I'm surprised you Freaks didn't think of it.
Here in Florida the major cities are almost entirely made for cars. There isn't a lot of public transportation (but when there is, it is often ineffective), not many pedestrian cross lights, and there are even vast areas with no sidewalks. What you do have in abundance are hit-and-run cases.
When you look around, the only people walking are those with extremely low incomes, who are incapable of paying for private transport. Those, I believe, are the majority of pedestrians being killed. The "killers" at the wheel, on the other hand, do have the money to pay for the car, the insurance, the lawyer, etc. And as you said, most Americans see it from a drive's perspective, because most Americans haven't experienced such poverty that would prevent them from owning a car.
I unfortunately have no data to back this, but it is an informed guess as I have been on the poor side of the road. I do believe it could be the reason why these cases are overlooked in places like Florida. Though I know in NYC the income distribution of pedestrians is more diverse than less metropolitan areas like Miami or Orlando, it could also be a factor.
In any case, I am personally a big proponent of public transportation and your latest podcast only makes me more angry that we have become so dependent on those machines (ie: cars). Why do people so completely ignore the amount of headaches and hardships they bring to our lives? You could make a whole podcast about that!



This is an issue discussed consistently in Urban Design. Cities have become increasingly auto oriented in the past 60 or so years and need to be rescaled for pedestrians. Removing curbs, creating bulb outs at street corners and changing the pavement texture at critical points all increase awareness for both the drivers and pedestrians . Simple infrastructure changes have been proven to have a huge impact in traffic calming and pedestrian safety. Streets are for the movement of people, who happen to be in cars. If you remove the need for the car through the walkable design the street becomes a safer place.

Glenn Mercer

Great podcast as always. My two cents? There is little movement to increasing criminal penalties for striking pedestrians because all of us ("us" defined as about 100,000,000 drivers in the USA) know it might be us who are charged next. We've all had near misses.

But also, while I agree every pedestrian death is "one too many," I was surprised that this economics podcast didn't look into the costs and benefits of reducing that number. There was some joking reference to wearing helmets, and an implicit view that criminal penalties should be increased, but by the end of the podcast I had no sense of how much we'd have to spend to reduce the number by 10% or 20%. Sure, Switzerland does a better job, but at what cost?