Drunk Driving: Is the Glass Half-Empty?

Last post, I gave you the good news: in 1982 there were 1.64 alcohol-related road fatalities per 100 million miles driven, and in 2007 there were only 0.43. There are a number of reasons for this terrific achievement: publicity, education, harsher penalties, stricter enforcement, and economic incentives.

Time to toast in celebration? Not quite. Because the curious fact is there’s a remarkable disconnect in our perceptions of DUI. Most of us are quite happy to believe that we have drunk driving under control and that further tough measures are unnecessary. Yet at the very same time we witness drunk driving laws being broken all the time – including by ourselves.

There are over 75,000 bars in the United States. Last time you were in one, did you really believe that every person in the room was getting home via cab or designated driver?

The last time you threw a party did you and your friends sit around drinking chocolate milk? If not, did everybody sleep over that night?

And let’s be honest: how many times have you climbed behind the wheel after a couple? I know, you’ve never been pulled over, which makes it tempting to think you’re not a drunk driver. But according to Paul Zador, Sheila Krawchuk, and B. Moore, the average person who is caught driving drunk has already gotten away with it 87 times. How about you?

I know, I know, you and your party guests weren’t really “drunk.” But despite the fact that it’s comforting to believe that all is dandy as long as you’re below the legal limit, E.J.D. Ogden and Herbert Moskowitz point out that the legal definition of DUI is fundamentally misleading.

Any amount of alcohol impairs you. According to one study, a blood alcohol concentration (BAC) of 0.05, which you attain after two drinks (very roughly, depending on your weight), and which is well below the legal limit, leaves a person with about a 38 percent increased risk of crashing. So, in the year 2000 alone, an estimated 2,600 people were killed in accidents involving drivers who were intoxicated but not technically under the influence. (See this from Dexter Taylor, Ted Miller, and Kenya Cox.)

(A third drink roughly multiplies the crash risk by about 2.7; a fourth by almost 5. Things skyrocket from there. See this from Richard D. Blomberg, Raymond C. Peck, Herbert Moskowitz, Marcelline Burns and Dary Fiorentino.)

According to the National Highway Traffic Safety Administration, a person is killed by a drunk driver on average every 45 minutes. A person is injured on average every minute.

According to Taylor et al., in 2000 only about one in every 140 miles driven in the U.S. was traversed by a legally drunk driver. Yet in 2007, 31.6 percent of America’s 41,059 traffic deaths involved at least one person with a BAC over the limit. About three in every 10 of us will be involved in an alcohol-related crash at some point during our lives.

And these figures are almost certainly understatements, since police undoubtedly fail to detect the alcohol in many cases.

In part due to these fatalities and injuries, drunk driving crashes cost us over $114 billion in the year 2000 alone. In today’s dollars, that’s about two-and-a-half times what it cost us to bail out GM.

And to make this worse, Mothers Against Drunk Driving reports that 63 percent of drunk driving’s costs are borne by others besides the intoxicated drivers. For example, we all pay for drunk driving through higher car insurance premiums.

Ted R. Miller, Rebecca S. Spicer and David T. Levy estimated that each mile driven by someone with a BAC over 0.08 costs society about $5.48 (1999 dollars) vs. about $0.11 for each sober mile. These costs include “medical care, public programs (police, fire, emergency medical and emergency transport), property damage, future earnings and lost quality of life.” Taylor et al. have written that each drink consumed costs us about one dollar due to the increase in alcohol-related crashes.

Lest you think this carnage is inevitable, when it comes to DUI the United States fares very poorly compared to other nations. The percentage of our fatal crashes which involve a drunk driver (as opposed to only sober parties) is higher than that of every other industrialized nation for which I’ve seen data, with the lone exception of Canada. As of 2004, the Japanese rate was about one-third of ours and the British rate about one-half. See this from the WHO and the World Bank.

(Then again, while we’re more likely to mix alcohol and driving, the Brits are far more likely to mix alcohol and watching soccer or going on holiday in Greece, which can also be pretty hazardous.)

The international comparisons show that we could be doing a lot more than we are, and last time I checked the Japanese and British civilizations hadn’t come crashing to the ground thanks to less drinking and driving. More could be done if we really had the willpower, and next time I’ll discuss some measures we could be taking. Then you can ask yourself the important question: how much am I really willing cut back on my partying to save lives?

(Addendum: Due to an editing mistake, an earlier version of this post incorrectly stated that “A third drink roughly multiplies the crash risk by about 2.7 percent; a fourth by almost 5 percent.”)

grad student

Interesting stuff, but neglects that in american sprawl cities, zoning laws often keep the bars and restaurants outside of walking distance from residential areas. Mass transit in such places is also rare to non-existent, as compared with britain and japan.


I don't really believe there is anything more we could do without tackling the dominance of car culture here. There simply aren't other options for most people to get home than the automobile, and even when they are many people are so car centric in their thinking they don't know they exist.

Want ot cut back on drunk driving? Improve the buses and convince people to take them.


I agree with all you say about drinking and driving, but you do not compare the one drink driver with the cell phone driver, the magazine reading driver, or the sandwich eating driver. I'm not defending the one drink driver, but let's not just tsk-tsk one risky habit.

Steve Rempel

I've often wondered about this, so I bought a cheap DOT approved breathalizer from Costco. Turns out a post dinner-out buzz for me is typically about 0.3. Lucky me. But this also showed me just how drunk and dangerous 0.8 really is.

Sean Samis

My father was a Highway Patrol man in a very rural state (SD) who spent a good deal of time scooping dead drunks off the road. Sometimes literally.

He drummed it into us at an early age that drunken driving is deadly.

My wife and I take turns being the DD, and we just don't drink and drive. It's not that hard to avoid, and most people who throw parties have other beverages and are supportive.


I think there's also a "disconnect" when you plainly see that drunk driving has been cut by almost 75% in the last generation...

...yet your argument for even stricter laws is that "you see it all the time."

And in comparing us to Europe, you forget the US' far more pervasive car culture. Transit in Europe is great, distances are small. We drive everywhere.

The flaws in logic throughout are incredible. I agree that no one should drive drunk, but you should be able to push that premise without sounding like you're inventing reasons why.


Classic moving-the-goalpost response from advocates who fear their goals are being achieved.

We've spent all this money over the last three decades to reduce DUIs, and what's the response? Spend even more money, to reduce it further.

Sorry, but you have yet to convince me of the cost-effectiveness of any further prevention and deterrence efforts.

And you have yet to contextualize prevalance and mortality data on DUIs in comparison to the effects of , say, red-light running, or lack of proper maintenance, or other road safety matters.

So feel free to advocate for your favorite pet cause. But as economists, I expect you to provide solid data backing up not only the prevalance and mortality rates of your chosen cause celebre, but also the cost-effectiveness of any solutions you might propose.

Eric H.

As an economist, you should know that laws will have unintended consequences. Current drunk driving laws likely have something of a moderating effect, such that people say "I will have only 2 drinks tonight since I have to drive home". If there is a zero tolerance law, this might change to all-or-none: "I will have 10 drinks tonight since I can't drive home". And there's a good chance people walk home rather than drive, which is even more dangerous. Does your calculation include the decrease in health/lifespan from the likely increase in drinking volume and drunk walking?


I agree with grad student. It seems inappropriate to conclude that the United States is doing less than it could be doing solely by making comparing fatal crash rates to that experienced by other western nations given the drastic differences in city layouts and transportation infrastructure. In fact, it is telling that the one country to rank worse in this statistic is Canada, the one country that is most similar to the United States in its dependence on transportation by private automobile and relative lack of effective public transportation options.


I can't believe the responses to this post. There is ABSOLUTELY no excuse EVER for driving under the influence of alcohol and yet reader after reader offers an excuse. Are they all THAT addicted to alcohol?


How about if we just take away the driving license forever of someone who gets a DUI? I bet that will encourage people not to drink and drive.


I live in Bay Ridge, a neighborhood in Brooklyn with strong Irish and Norwegian roots. We have bars everywhere. Walk along 3rd or 5th avenue, and it would be tough to go 2 blocks without finding a bar. So its easy for someone there to have a good time drinking and not worry about how they will get home. The big problems come mainly from Staten Islanders who come over the Verrazano to drink here.


Frank, you are making a mistake by attributing bad behavior to commenters who are trying to tackle a difficult reality. There is no excuse for drinking and driving but people do it anyway. If everyone were conscientious and responsible all the time then we wouldn't need anti-drunk driving programs in the first place, people would have never been doing it. Heck, if everyone were responsible we'd hardly need laws or government at all. But they aren't so we need to explain what the problem is and suggest solutions.

Personally, I live within walking distance of anywhere I normally go and if I want to go out somewhere further away I will alternate with my girlfriend on who is designated driver. Alcoholism is not a determinant of my actions but I do understand that just because I choose to act this way that there is no way to force others to conform to my ideal of personal responsibility. This means it is worthwhile speculating on the cause of why we see increased drunk driving fatalities on the US and the effectiveness of various possibilities of dealing with the problem.



"A third drink roughly multiplies the crash risk by about 2.7 percent"

So having a third drink reduces my crash risk by 97.3%? Awesome!


Does the research split it up based on different parts of the U.S.? I would imagine places like NYC, Boston and Chicago with lots of mass transit options would have lower deaths from DUIs per capita than places like Miami which has nearly no major mass transit (The Metrorail is the one main mass transit but doesn't run 24 hours a day and few homes are close to any of the stops).

@ Frank

I don't see any of the comments excusing the act of drunk driving. People are offering opinions as to why the numbers are different in the US than most other countries. EVM has it correct, the one main country that has a worse drunk driving rate than the US is Canada, which has a transportation infrastructure similar to ours.


I have never seen an article more logically and mathematically dishonest and ignorant than this post, for the following reasons:

1. No person should be even be listened to who uses "alcohol-related" in any serious discussion of risk evaluation. Every single road fatality is "water-related." What the OP is looking for is "alcohol-caused" which can only be determined in controlled scientific experiment or in a court of law. Anything else is Mad Mother nonsense.

2. It has already been shown on this very blog that the rational actor leaving a bar drunk is at greater risk of death if he were to walk home instead of drive. Are Mad Mothers offering free rides home to single drinkers? I don't think so.

3. Anything but staying at home puts you and everybody else at risk. Thus, the logical result of the OP argument is that nobody should ever leave home. Mad Mothers would probably approve. Furthermore, alcohol "impairs" sex, homework and conversation, putting "social costs" on the rest of us. Do Mad Mothers want to prohibit drinking before sex, drinking while doing homework, drinking while conversing? Maybe. Some of us find eating, sex, homework and conversing tolerable ONLY after drinking.

4. If "a person is killed by a drunk driver on average every 45 minutes," the simple solution is to find that driver and put him away, not bother the rest of us happy drinkers.

5. Statements such as "drunk driving crashes cost us over $114 billion in the year 2000 alone. In today's dollars, that's about two-and-a-half times what it cost us to bail out GM" are PURE NONSENSE, absent proof of causation (see point 1 above).

6. The Mad Mothers' assertion that "63 percent of drunk driving's costs are borne by others" is of questionable interest. Much the same figure applies to child rearing, gun battles, drug criminalization, national park visits, and so on. Does that mean we should privatize education, eliminate kiddy tax deductions, outlaw gun possession, legalize drugs, sell off the national parks, etc? Indeed, these are far better ideas than attempting a Prohibition II.

7. The statement that "the last time I checked the Japanese and British civilizations hadn't come crashing to the ground thanks to less drinking and driving" is stupid as hell. First of all, our civilization is not threatening to come crashing to the ground now as much as it was during Prohibition. Secondly, we used to be able to drink while driving in Texas, with the result that a guy would get off work, stop and buy a six-pack, and drive home drinking a single beer. Now, what he does is stop at a bar near work, drink two or three, and then drive home. This results in collateral damage caused by Mad Mother laws. Furthermore, in Tokyo, you can use coins and bills, credit cards and cellphones to get a liter of whisky or a cold beer from a vending machine right on the street. Even if you are drunk. Even if you are 12 years old. Take off your blinders, Eric Morris.

7. Whether insurance premiums go up or not is irrelevant as well. People need not carry insurance. Wisconsin and New Hampshire are enlightened enough to see that, and anyone can drive without insurance if he doesn't like paying premiums. Of course insurance premiums rise. They rise according to sex, age, driving record, marital status, family status, credit status, state of domicile and for all kinds of illogical reasons.

8. Someone needs to inform Ted R. Miller, Rebecca S. Spicer and David T. Levy that society forces these costs on folks by its socialist insurance and medical care programs. Can someone who likes to drink opt the hell out of these socialist programs? If somebody feels injured by a drunk driver, he can sue. That's the way Amerika works. If somebody sober negligently hits you with his car, you can sue. Do Mad Mothers also want to outlaw driving while sober?

9. The OP says "The percentage of our fatal crashes which involve a drunk driver (as opposed to only sober parties) is higher than that of every other industrialized nation for which I've seen data, with the lone exception of Canada." I challenge him to show these data. They DO NOT exist. (see point 1 above). The study he cites talks about alcohol use as a "factor" in the accident. As a scientist, I consider "factor" to be limited to a factor in causation-it cannot mean an incidental occurrence. The fact that the Twin Towers fell on a clear day does not lead to the conclusion that clear weather was a factor in the collapse. Not even if all the towers in history since Sampson's attack occurred in clear weather. There is a huge gulf between science and this blog, dear reader.



I think "grad student" brings up an excellent point, although I'm also disappointed that Freakonomics hasn't already considered the affects of the National Minimum Drinking Age Act, passed in 1984 (http://en.wikipedia.org/wiki/National_Minimum_Drinking_Age_Act). This legislation was designed less because of concern about 18-20 year olds, and arguably because they were easy prey for legislation that could theoretically reduce the number of drinkers on the road. Voila! It seems to have met that objective.

Eileen Wyatt

I'm puzzled as to why readers are told we've broken drunk driving laws by driving with a blood alcohol level under the legal limit. This does not break the law.

If your data showing greater harms from people driving over the legal limit is accurate, then it is also not an impairment in any practical sense, as if it were, the deaths and injuries from "non-legally-drunk" drivers would be closer to those from legally drunk drivers.

If we're going to assume (contra the data) that a drink with dinner makes a normal, healthy person too impaired to drive safely, what about the stronger forms of allergy or cold medication? What about being tired? Or in a bad mood?

Given the data, surely catching a greater percentage of legally drunk drivers (or preventing them from hitting the road) would be more helpful than hassling people who are not particularly dangerous but have had a drink.



The wife took umbrage to this last night. Upset that Superfreakonomics only gave two options of walking drunk or driving that way. I explained that she was participating an a game of Ultimatum, but then remembered that she'd not gotten that far in the book.

On another driving issue, I was talking to a vetran over the road truck driver today about the impact of GPS on his profession. I was suprised to lear it was a nightmare. While helping greener drivers with some routes, the rookies were also following the directions to a fault. Ending up the victims of low bridges and off-road excursions.

Has anyone seen any data to back this up. I'd love to explore the idea further.


Two drinks puts you over the legal limit of 0.08?
What does sleep deprivation do? Are you getting your full eight hours a night?
What about being on a cell phone?
What about crying kids in the back seat?

And, of course, what about time? I often have a beer with dinner at a restaurant. By the time we finish desert, that beer ought to be well metabolized and gone.

(On the other hand, when the proprietor of a local steakhouse offered to "buy" each of us a port after a great recent meal, I told my wife, "You're Driving" and happily drank both glasses. As I said, it's all about time making the booze go away.)