terms of traffic safety: Based on statistics from the International Traffic Safety Data and a.n.a.lysis Group; retrieved on January 13, 2007, from http://cemt.org/IRTAD/IRTADPUBLIC/we2.htm.

some 160 deaths per 10,000 vehicles: World Report on Road Traffic Injury Prevention World Report on Road Traffic Injury Prevention (Geneva: World Health Organization and World Bank, April 4, 2004). (Geneva: World Health Organization and World Bank, April 4, 2004).

"to use the buses": BBC, February 28, 2001. Accessed from: http://news.bbc.co.uk/2/hi/africa/1186572.stm.

slightly higher in Belgium): Pocket World in Figures 2007 Pocket World in Figures 2007 (London: Economist, 2007). (London: Economist, 2007).

risk of traffic fatalities: See Theodore E. Keeler, "Highway Safety, Economic Behavior, and Driving Environment," American Economic Review, American Economic Review, vol. 84, no. 3 (1994), pp. 68493, and Reid Ewing, Richard A. Schieber, and Charles V. Zegeer, "Urban Sprawl as a Risk Factor in Motor Vehicle Occupant and Pedestrian Fatalities," vol. 84, no. 3 (1994), pp. 68493, and Reid Ewing, Richard A. Schieber, and Charles V. Zegeer, "Urban Sprawl as a Risk Factor in Motor Vehicle Occupant and Pedestrian Fatalities," American Journal of Public Health, American Journal of Public Health, vol. 93, no. 9 (2003), pp. 154145. vol. 93, no. 9 (2003), pp. 154145.

Belgium had 522: World Report on Road Traffic Injury Prevention, World Report on Road Traffic Injury Prevention, op. cit., p. 198. op. cit., p. 198.

fairness of the process: Tom R. Tyler, Why People Obey the Law Why People Obey the Law (Princeton, N.J.: Princeton University Press, 2006). (Princeton, N.J.: Princeton University Press, 2006).

The information on Belgium's traffic enforcement comes from Lode Vereeck and Lieber Deben, "An International Comparison of the Effectiveness of Traffic Safety Enforcement Policies," unpublished paper, Limburg University, Belgium, 2003.

lowest crash rates in the world: Retrieved from the International Road Traffic and Accident Database (IRTAD), at http://cemt.org/IRTAD/IRTADPUBLIC/we2.htm.

after-tax income: Before 1999, fines were based on pre pretax income, says Heikki Summala of the Traffic Research Unit at the University of Helsinki. This means fines have dropped between 20 and 60 percent, but at the same time minimum fines were raised, so revenue has in fact increased. E-mail correspondence with Heikki Summala, November 9, 2007.

Jaakko Rytsola: The Finnish speeding ticket information comes from Steve Stecklow, "Finnish Drivers Don't Mind Sliding Scale, but Instant Calculation Gets Low Marks," Wall Street Journal, Wall Street Journal, January 2, 2001. January 2, 2001.

return to shortly: A Finnish public-opinion poll in 2001 found that 66 percent of male drivers and 73 percent of male nondrivers felt the fine system was fair, while 77 percent of female drivers and 78 percent of female nondrivers thought it fair. The data comes from a study (in Finnish): T. Lappi-Seppala, "Public Opinion and the 1999 Reform of the Day-Fine System," National Research Inst.i.tute of Legal Policy, Publication No. 195, Helsinki, 2002. Thanks to Heikki Summala for providing the numbers.

rather stagnant: In 2003, for example, according to Eurostat, it grew just .50 percent. Data obtained from http://epp.eurostat.ec.europa.eu. Had the GDP risen, there may have been an increase in fatalities, reflecting the higher amounts of driving due to economic vitality-but it certainly would not have been by enough to offset the huge reductions in fatalities.

been in a crash): E. Lagarde, M. Chiron, and S. Lafont. "Traffic Ticket Fixing and Driving Behaviours in a Large French Working Population," Journal of Epidemiology and Community Health, Journal of Epidemiology and Community Health, vol. 58 (2004), pp. 56268. vol. 58 (2004), pp. 56268.

hundreds of traffic fatalities: Alexandre Dorozynski, "French Elections Can Kill," British Medical Journal, British Medical Journal, November 3, 2001, p. 1021. November 3, 2001, p. 1021.

The lesson is: At least one a.n.a.lysis posits that income equality is related in a linear fashion to traffic fatalities; e.g., in both poor, and, to a lesser extent, wealthy countries, the traffic fatality rate may be affected by the level of income equality. Perhaps not surprisingly, the Scandinavian countries, among the leaders in income equality, also rank near the top in traffic safety. See Nejat Anbarci, Monica Escaleras, and Charles Register, "Income, Income Inequality and the 'Hidden Epidemic' of Traffic Fatalities," No. 5002, Working Papers from Department of Economics, College of Business, Florida Atlantic University. Retrieved from http://econpapers.repec.org/paper/falwpaper/05002.htm.

and traffic fatalities: This relationship is argued in, among other sources, D. Treisman, "The Causes of Corruption: A Cross-National Study," traffic fatalities: This relationship is argued in, among other sources, D. Treisman, "The Causes of Corruption: A Cross-National Study," Journal of Public Economics, Journal of Public Economics, no. 76 (June 2000), pp. 399457. no. 76 (June 2000), pp. 399457.

income and traffic fatalities: See Nejat Anbarci, Monica Escaleras, Monica Register, and Charles A. Register, "Traffic Fatalities and Public Sector Corruption," Kyklos, Kyklos, vol. 59, no. 3 (August 2006), pp. 32744; available at vol. 59, no. 3 (August 2006), pp. 32744; available at http://ssrn.com/abstract=914243.

of Europe's road fatalities: See "Fools and Bad Roads," Economist, Economist, March 22, 2007. March 22, 2007.

rewards inefficient firms: For a good review of the various debates over corruption and growth, see P. Bardhan, "Corruption and Development: A Review of Issues," Journal of Economic Literature, Journal of Economic Literature, vol. 35 (September 1997), pp. 132046. vol. 35 (September 1997), pp. 132046.

beneath the acceptable "minimum": See Daniel Kaufmann, "Corruption: The Facts," Foreign Policy, Foreign Policy, no. 107 (Summer 1997), pp. 11431. no. 107 (Summer 1997), pp. 11431.

because of corruption: The most extreme case of this may be Lagos, the largest city in Nigeria and predicted to be among the world's largest cities in the next decade. The average commuter in Lagos is said to face myriad challenges. These begin with the crumbling roads and infrastructure, which have scarcely been repaired since being erected in the oil boom of the 1970s; they themselves are a kind of symbol of the endemic corruption of Nigeria, where close to $400 billion in oil revenues were sequestered out of the country in a forty-year period. Other challenges include arbitrary fees charged at will by bus drivers and their quasi-official a.s.sociates, the of corruption: The most extreme case of this may be Lagos, the largest city in Nigeria and predicted to be among the world's largest cities in the next decade. The average commuter in Lagos is said to face myriad challenges. These begin with the crumbling roads and infrastructure, which have scarcely been repaired since being erected in the oil boom of the 1970s; they themselves are a kind of symbol of the endemic corruption of Nigeria, where close to $400 billion in oil revenues were sequestered out of the country in a forty-year period. Other challenges include arbitrary fees charged at will by bus drivers and their quasi-official a.s.sociates, the agberos, agberos, not to mention the numerous unofficial roadblocks, manned by gangs of unemployed "area boys," that drivers must navigate. The multiple levels of corruption present in-and contributing to-Lagos's epic "go-slows" were demonstrated in an astonishing story told by the journalist George Packer. While riding on the streets of Lagos, Packer's driver was stopped by an not to mention the numerous unofficial roadblocks, manned by gangs of unemployed "area boys," that drivers must navigate. The multiple levels of corruption present in-and contributing to-Lagos's epic "go-slows" were demonstrated in an astonishing story told by the journalist George Packer. While riding on the streets of Lagos, Packer's driver was stopped by an agbero, agbero, who demanded money to help the driver negotiate who demanded money to help the driver negotiate another another bribe, with the official traffic police. The traffic cop intervened, if only to collect the bribe-not doing so, it seemed, would actually make the police officer look as if he were derelict in his duty. See George Packer, "The Megacity: Decoding the Chaos of Lagos," bribe, with the official traffic police. The traffic cop intervened, if only to collect the bribe-not doing so, it seemed, would actually make the police officer look as if he were derelict in his duty. See George Packer, "The Megacity: Decoding the Chaos of Lagos," New Yorker, New Yorker, November 26, 2006. See also Adewale Ajayi, November 26, 2006. See also Adewale Ajayi, Nigerian Tribune, Nigerian Tribune, March 2, 2007; and Osise Dibosa, "Olubunmi Peters and Ferma," March 2, 2007; and Osise Dibosa, "Olubunmi Peters and Ferma," This Day, This Day, June 12, 2007. June 12, 2007.

take their place: Benjamin A. Olken and Patrick Barron, "The Simple Economics of Extortion: Evidence from Trucking in Aceh," NBER Working Paper No. 13145, National Bureau of Economic Research, June 2007.

"work repairing potholes": Robert Guest, "The Road to h.e.l.l Is Unpaved," Economist, Economist, December 19, 2002. December 19, 2002.

"actual driving skill": The Delhi driving-license experiment is detailed in Marianne Bertrand, Simeon Djankov, Rema Hanna, and Sendhil Mullainathan, "Does Corruption Produce Unsafe Drivers?" NBER Working Paper No. 12274, National Bureau of Economic Research, June 2006.

"clarity of purpose": This line comes from Pavan K. Varma, Being Indian Being Indian (London: Penguin Books, 2005), p. 79. (London: Penguin Books, 2005), p. 79.

some 150,000 tickets: Raymond J. Fisman and Edward Miguel, "Cultures of Corruption: Evidence from Diplomatic Parking Tickets," NBER Working Paper No. W12312 (June 2006). Retrieved at http://ssrn.com/abstract=910844.

the city of London: Retrieved from Channel Four News Online, Channel Four News Online, http://www.channel4.com/news/articles/society/environment/diplomatic+ccharge+bill+ tops+45m/569892.

pays the charge: Nicola Woolc.o.c.k, "Nations Unite to Join a Boycott of Congestion Charge," Times Times (London), February 21, 2007. (London), February 21, 2007.

norms regarding them: This is why we can often see compliance with traffic laws differing even within within a country. In Italy, corruption is more endemic in the south than the north, for reasons, as mentioned in an earlier note, having to do with varying degrees of civic culture. And so as the state seems to gradually wither away the farther south you go, so too does the traffic behavior come to have less to do with the law. In 2000, a national helmet law was pa.s.sed for motorcyclists of any age. Afterward, usage rates in the north were reported as high as 95 percent. In the south, however, they were only as high as 70 percent, and as low as 50 percent. For corruption levels, see Alfredo del Monte and Erasmo Papagni, "The Determinants of Corruption in Italy: Regional Panel Data a.n.a.lysis," a country. In Italy, corruption is more endemic in the south than the north, for reasons, as mentioned in an earlier note, having to do with varying degrees of civic culture. And so as the state seems to gradually wither away the farther south you go, so too does the traffic behavior come to have less to do with the law. In 2000, a national helmet law was pa.s.sed for motorcyclists of any age. Afterward, usage rates in the north were reported as high as 95 percent. In the south, however, they were only as high as 70 percent, and as low as 50 percent. For corruption levels, see Alfredo del Monte and Erasmo Papagni, "The Determinants of Corruption in Italy: Regional Panel Data a.n.a.lysis," European Journal of Political Economy, European Journal of Political Economy, vol. 23 (June 2007), pp. 37996. For helmet-use rates, see F. Servadei, C. Begliomini, E. Gardini, M. Giustini, F. Taggi, and J. Kraus, "Effect of Italy's Motorcycle Helmet Law on Traumatic Brain Injuries," vol. 23 (June 2007), pp. 37996. For helmet-use rates, see F. Servadei, C. Begliomini, E. Gardini, M. Giustini, F. Taggi, and J. Kraus, "Effect of Italy's Motorcycle Helmet Law on Traumatic Brain Injuries," Injury Prevention, Injury Prevention, vol. 9, no. 3 (2003), pp. 25760. vol. 9, no. 3 (2003), pp. 25760.

casualties there will be: See D. Parker, J. T. Reason, A. S. R. Manstead, and S. G. Stradling, "Driving Errors, Driving Violations and Accident Involvement," Ergonomics, Ergonomics, vol. 38 (1995), pp. 103648. vol. 38 (1995), pp. 103648.

more women in government: Anand Swamy, Stephen Knack, Young Lee, and Omar Azfar, "Gender and Corruption," Center for Development Economics, Department of Economics, Williams College, 2000.

Chapter Nine: Why You Shouldn't Drive with a Beer-Drinking Lawyer our brains as we drive: Research has shown that the various aspects of driving, everything from following a traffic rule (e.g., specifying a one-way street) to navigating a set of directions to antic.i.p.ating the actions of other drivers, seem to trigger discrete activity in a variety of brain regions and networks. Researchers at University College London, for example, have monitored drivers as they "drove" the detailed recreation of London found in the popular video game The Getaway. The Getaway. See H. J. Spiers and E. A. Maguire, "Neural Substrates of Driving Behaviour," See H. J. Spiers and E. A. Maguire, "Neural Substrates of Driving Behaviour," NeuroImage, NeuroImage, vol. 36 (2007), pp. 24555. vol. 36 (2007), pp. 24555.

fifty thousand times a year: P. G. Martin and A. L. Burgett, "Rear-End Collision Events: Characterization of Impending Crashes," Proceedings of the First Human-Centered Transportation Simulation Conference Proceedings of the First Human-Centered Transportation Simulation Conference (Iowa City: University of Iowa, 2000). (Iowa City: University of Iowa, 2000).

walks away alive: See Jack Stuster, "The Unsafe Driving Acts of Motorists in the Vicinity of Large Trucks," U.S. Department of Transportation, Federal Highway Administration, Office of Motor Carriers and Highway Safety, February 1999.

should probably fear: See L. J. Armony, D. Servan-Schreiber, J. D. Cohen, and J. E. LeDoux, "An Anatomically-Constrained Neural Network Model of Fear Conditioning," Behavioral Neurocience, Behavioral Neurocience, vol. 109 (1995), pp. 24656. vol. 109 (1995), pp. 24656.

dangerous nature of trucks: Opinion surveys of car drivers tend to find mostly negative opinions of truck drivers' behavior. See, for example, Robert S. Moore, Stephen LeMay, Melissa L. Moore, Pearson Lidell, Brian Kinard, and David McMillen, "An Investigation of Motorists' Perceptions of Trucks on the Highways," Transportation Journal, Transportation Journal, January 5, 2001. January 5, 2001.

responsibility in the crash: Daniel Blower, "The Relative Contribution of Truck Drivers and Pa.s.senger Vehicles to Truck-Pa.s.senger Vehicle Traffic Crashes," report prepared for the U.S. Department of Transportation, Federal Highway Administration, Office of Motor Carriers, June 1998.

is actually the case): This may be the "availability heuristic" at work again. Large trucks, in part because they are driven longer distances and tend to be on the road at the same time as most motorists, seem to be more prevalent than they really are. A Canadian study found that while motorists believed that the number of trucks on the roads was rising, the number actually dropped dropped during the period in question (while the number of cars grew). See Gordon G. Baldwin, "Too Many Trucks on the Road?" Transportation Division, Statistics Canada, Ottawa. during the period in question (while the number of cars grew). See Gordon G. Baldwin, "Too Many Trucks on the Road?" Transportation Division, Statistics Canada, Ottawa.

"risk as a.n.a.lysis": Paul Slovic, Melissa L. Finucane, Ellen Peters, and Donald G. MacGregor, "Risk as a.n.a.lysis and Risk as Feelings: Some Thoughts About Affect, Reason, Risk, and Rationality," Risk a.n.a.lysis, Risk a.n.a.lysis, vol. 24, no. 2 (2004), pp. 31123. vol. 24, no. 2 (2004), pp. 31123.

50 years of driving: Data retrieved on May 5, 2007, from http://hazmat.dot.gov/riskrngmt/riskcompare.htm.

the lifetime probability: P. Slovic, B. Fischhoff, and S. Lichtenstein, "Accident Probabilities and Seat Belt Usage: A Psychological Perspective," Accident a.n.a.lysis & Prevention, Accident a.n.a.lysis & Prevention, vol. 13 (1978), pp. 28185. vol. 13 (1978), pp. 28185.

"the danger of leaving home": William H. Lucy, "Mortality Risk a.s.sociated with Leaving Home: Recognizing the Relevance of the Built Environment," American Journal of Public Health, American Journal of Public Health, vol. 93, no. 9 (September 2003), pp. 156469. vol. 93, no. 9 (September 2003), pp. 156469.

eleven times that: This figure was provided to me by Per Garder, a professor of civil and environmental engineering at the University of Maine. Using the required risk exposure levels as quoted by the Occupational Safety and Health Administration (in "Occupational Exposure to Asbestos," Federal Register 59:40964-41161, 1994, and OSHA Preambles, "Blood Borne Pathogens," 29 CFR 1910.1030, Federal Register 56:64004, 1991: 29206), Garder notes that the risk of dying over a lifetime in manufacturing and service employment, respectively, "must be less than 1. 8 and 1.0 deaths per 1,000 employees." By those standards, Garder extrapolates if 1 person in a 1,000 were "allowed" to die in traffic over an average of 77 years of life, 1 person in 77,000 would thus be allowed to die in America this year in a traffic accident. Using America's population of 300 million, 1 in 77,000 would be 3,896 people. But the fatality figure was over 11 times that. In other words, if traffic were an industry-whether heavy manufacturing or service-it would have been shut down a long time ago.

every thirty-two minutes: Fatality statistics were taken from Traffic Safety Facts 2004 Traffic Safety Facts 2004 (Washington, D.C.: National Highway Traffic Safety Administration, 2005). (Washington, D.C.: National Highway Traffic Safety Administration, 2005).

3 out of every 1,000: Clifford Winston, Vikram Maheshri, and Fred Mannering, "An Exploration of the Offset Hypothesis Using Disaggregate Data: The Case of Airbags and Antilock Brakes," Journal of Risk Uncertainty, Journal of Risk Uncertainty, vol. 32 (2006), pp. 8399. vol. 32 (2006), pp. 8399.

raises the crash risk: M. G. Lenne, T. J. Triggs, and J. R. Redman, "Time of Day Variations in Driving Performance," Accident a.n.a.lysis & Prevention, Accident a.n.a.lysis & Prevention, vol. 29, no. 4 (1997), pp. 43137, and G. Mayc.o.c.k, "Sleepiness and Driving: The Experience of U.K. Car Drivers," vol. 29, no. 4 (1997), pp. 43137, and G. Mayc.o.c.k, "Sleepiness and Driving: The Experience of U.K. Car Drivers," Accident a.n.a.lysis & Prevention, Accident a.n.a.lysis & Prevention, vol. 29, no. 4 (1997), pp. 45362. vol. 29, no. 4 (1997), pp. 45362.

day to be on the road: As David Klein and Julian Waller noted, the posting of holiday traffic fatalities presents several problems. "Although absolute numbers may serve a purpose in indicating the raw impact of highway crashes on the nation or on a community," they write, "their use provides only a partial indication of magnitude and often a misleading indication of trends. First, fatality figures ignore the 1. 5 to 3 million annual non-fatal injuries-which may represent a social cost far higher than the 56,000 fatalities. Second, the 'holiday death toll' may give drivers an unjustified feeling of anxiety on holiday weekends and a false sense of security on weekdays if it persuades them that the holiday incidence is substantially higher than on weekdays." From Klein and Waller, "Causation, Culpability and Deterrence in Highway Crashes," prepared for the Department of Transportation, July 1970, p. 27.

week before or after: C. M. Farmer and A. F. Williams. "Temporal Factors in Motor Vehicle Crash Deaths," Injury Prevention, Injury Prevention, vol. 2 (2005), pp. 1823. vol. 2 (2005), pp. 1823.

should be about $8,000: Steven D. Levitt and Jack Porter, "How Dangerous Are Drinking Drivers?," Journal of Political Economy, Journal of Political Economy, vol. 109, no. 6 (2001), pp. 11981237. The authors rely on a clever statistical trick that does not require knowing the actual number of drinking and sober drivers on the road (a number that would be extremely hard to come by in any case) but, rather, uses an extrapolation taken from the relative proportion of sober and drunk drivers involved in two-car crashes. Levitt and Porter generate their relative risk numbers by looking at two-car crashes and "the relative frequency of accidents involving two drinking drivers, two sober drivers, or one of each." This information, they argue, "is sufficient to separately identify both the relative likelihood of causing a fatal crash on the part of drunk and sober drivers and the fraction of drivers on the road who have been drinking." vol. 109, no. 6 (2001), pp. 11981237. The authors rely on a clever statistical trick that does not require knowing the actual number of drinking and sober drivers on the road (a number that would be extremely hard to come by in any case) but, rather, uses an extrapolation taken from the relative proportion of sober and drunk drivers involved in two-car crashes. Levitt and Porter generate their relative risk numbers by looking at two-car crashes and "the relative frequency of accidents involving two drinking drivers, two sober drivers, or one of each." This information, they argue, "is sufficient to separately identify both the relative likelihood of causing a fatal crash on the part of drunk and sober drivers and the fraction of drivers on the road who have been drinking."

doubling of the speed: H. C. Joksch, "Velocity Change and Fatality Risk in a Crash: A Rule of Thumb," Accident a.n.a.lysis & Prevention, Accident a.n.a.lysis & Prevention, vol. 25, no. 1 (1993), pp. 10304. vol. 25, no. 1 (1993), pp. 10304.

doing 30 miles per hour: Allan F. Williams, Sergey Y. Krychenko, and Richard A. Retting, "Characteristics of Speeders," Journal of Safety Research, Journal of Safety Research, vol. 37 (2006), pp. 22732. vol. 37 (2006), pp. 22732.

get into more crashes: See, for example, Williams, Kyrychenko, and Retting. "Characteristics of Speeders," ibid.

additional 5 kilometers per hour: See C. N. Kloeden, A. J. McLean, V. M. Moore, and G. Ponte, "Travelling Speed and the Risk of Crash Involve ment," NHMRC Road Accident Research Unit, University of Adelaide, November 1997.

"relatively high speed drivers": David Solomon, Accidents on Main Rural Highways Related to Speed, Driver, and Vehicle Accidents on Main Rural Highways Related to Speed, Driver, and Vehicle (Washington, D.C.: U.S. Department of Commerce, Bureau of Public Roads, 1964). (Washington, D.C.: U.S. Department of Commerce, Bureau of Public Roads, 1964).

flow in smooth harmony: The speed-variance argument was most famously taken up by Charles Lave, "Speeding, Coordination, and the 55 MPH Limit," American Economic Review, American Economic Review, vol. 75, no. 5 (December 1985), pp. 115964. Interestingly, in a point that has not been emphasized by those later citing Lave, he writes: "Although I have found no statistically discernible effect from speed, per se, this does not necessarily imply that it is safe to raise the speed limit, for we do not know what effect a higher limit would have on the speed variance." If the speed limit is 65 miles per hour but many people are driving 75, it does not necessarily follow that raising it to 75 miles per hour will reduce speed variance or make things safer. Do we want the drivers who feel comfortable at a lower level forced to go faster? Do we vol. 75, no. 5 (December 1985), pp. 115964. Interestingly, in a point that has not been emphasized by those later citing Lave, he writes: "Although I have found no statistically discernible effect from speed, per se, this does not necessarily imply that it is safe to raise the speed limit, for we do not know what effect a higher limit would have on the speed variance." If the speed limit is 65 miles per hour but many people are driving 75, it does not necessarily follow that raising it to 75 miles per hour will reduce speed variance or make things safer. Do we want the drivers who feel comfortable at a lower level forced to go faster? Do we want want Grandma and Grandpa driving 75 miles per hour? Grandma and Grandpa driving 75 miles per hour?

held by young males: T. Horberry, L. Hartley, K. Gobetti, F. Walker, B. Johnson, S. Gersbach, and J. Ludlow, "Speed Choice by Drivers: The Issue of Driving Too Slowly," Ergonomics, Ergonomics, vol. 47, no. 14 (November 2004), pp. 156170. vol. 47, no. 14 (November 2004), pp. 156170.

at low speeds: For elaboration on this point, see Kloeden, McLean, Moore, and Ponte, "Travelling Speed," op. cit.

involved a stopped vehicle: Ronald K. Knipling, "IVHS Technologies Applied to Collision Avoidance: Perspectives on Six Target Crash Types and Countermeasures," technical paper presented at the Safety and Human Factors session of 1993 IVHS America Annual Meeting, Washington, D.C., April 1417, 1993.

not hold for individuals: Gary A. Davis, "Is the Claim That 'Variance Kills' an Ecological Fallacy?," Accident a.n.a.lysis & Prevention, Accident a.n.a.lysis & Prevention, vol. 34 (2002), pp. 34346. With the Solomon curve, Davis argues that one cannot determine the individual driver's crash risk by looking at the whole. Solomon's curve, maintains Davis, is a purely mathematical effect that says little about how the world works, "like saying an object is heavy because it weighs more." Another problem with the Solomon curve is that it does not explain causes. If twenty cars slowing for traffic congestion-and thus going below the median speed-were struck by ten cars traveling at the median and ten cars traveling above the median, the resulting "curve" would indeed suggest that slower drivers were the most at risk of being in a crash. But looking at each crash individually, one would conclude that the faster-moving cars had actually been the source of the risk for the slower-moving cars. As an example of a ecological fallacy, the statistician David Freedman has compared the income levels of U.S. states against the percentage of foreign-born residents in each. Doing this, one could make a statistically robust "correlation" that says foreign-born residents of the United States earn more than native-born residents, when actually the vol. 34 (2002), pp. 34346. With the Solomon curve, Davis argues that one cannot determine the individual driver's crash risk by looking at the whole. Solomon's curve, maintains Davis, is a purely mathematical effect that says little about how the world works, "like saying an object is heavy because it weighs more." Another problem with the Solomon curve is that it does not explain causes. If twenty cars slowing for traffic congestion-and thus going below the median speed-were struck by ten cars traveling at the median and ten cars traveling above the median, the resulting "curve" would indeed suggest that slower drivers were the most at risk of being in a crash. But looking at each crash individually, one would conclude that the faster-moving cars had actually been the source of the risk for the slower-moving cars. As an example of a ecological fallacy, the statistician David Freedman has compared the income levels of U.S. states against the percentage of foreign-born residents in each. Doing this, one could make a statistically robust "correlation" that says foreign-born residents of the United States earn more than native-born residents, when actually the opposite opposite is true. See David A. Freedman, "Ecological Inference and the Ecological Fallacy," in is true. See David A. Freedman, "Ecological Inference and the Ecological Fallacy," in International Encyclopedia of the Social & Behavioral Sciences, International Encyclopedia of the Social & Behavioral Sciences, vol. 6, ed. N. J. Smelser and Paul B. Baltes (New York: Pergamom, 2001), pp. 402730. vol. 6, ed. N. J. Smelser and Paul B. Baltes (New York: Pergamom, 2001), pp. 402730.

in the same direction: E. C. Cerrelli, "1996 Traffic Crashes, Injuries, and Fatalities-Preliminary Report," Report No. DOT HS 808 543, National Highway Traffic Safety Administration, March 1997. I was alerted to this finding by an excellent report summarizing the various speed issues. See Jack Stuster and Zail Coffman (1998), Synthesis of Safety Research Related to Speed and Speed Limits, Synthesis of Safety Research Related to Speed and Speed Limits, FHWA-RD-98-154 (Washington, D.C.: Federal Highway Administration, 1998). FHWA-RD-98-154 (Washington, D.C.: Federal Highway Administration, 1998).

whose teams had lost: D. A. Redelmeier and C. L. Stewart, "Do Fatal Crashes Increase Following a Super Bowl Telecast?" Chance, Chance, vol. 18, no. 1 (2005), pp. 1924. vol. 18, no. 1 (2005), pp. 1924.

have been drinking: R. G. Smart, "Behavioral and Social Consequences Related to the Consumption of Different Beverage Types," Journal of Studies on Alcohol, Journal of Studies on Alcohol, vol. 57 (1996), pp. 7784. vol. 57 (1996), pp. 7784.

at .08 to .1 percent: R. P. Compton, R. D. Blomberg, H. Moskowitz, M. Burns, R. C. Peck, and D. Fiorentino, "Crash Risk of Alcohol Impaired Driving," Proceedings of the 16th International Conference on Alcohol, Drugs and Traffic Safety, Proceedings of the 16th International Conference on Alcohol, Drugs and Traffic Safety, CD-ROM (Montreal, Societe de l'a.s.surance Automobile du Quebec, 2002). CD-ROM (Montreal, Societe de l'a.s.surance Automobile du Quebec, 2002).

BAC of zero: R. F. Borkenstein, R. F. Crowther, R. P. Shumate, W. B. Ziel, and R. Zylman, "The Role of the Drinking Driver in Traffic Accidents," Bloomington, Indiana, Department of Police Administration and Indiana University, 1964.

"handling" a small intake: See, for example, Leonard Evans, Traffic Safety Traffic Safety (Bloomfield Hills: Science Serving Society, 2004), p. 246. (Bloomfield Hills: Science Serving Society, 2004), p. 246.

shown up in other studies: P. L. Zador, S. A. Krawchuk, and R. B. Voas, Relative Risk of Fatal and Crash Involvement by BAC, Age and Gender Relative Risk of Fatal and Crash Involvement by BAC, Age and Gender (Rockville, Md.: Westat, April 2000). (Rockville, Md.: Westat, April 2000).

statistically less safe: Paul M. Hurst, David Harte, and William Frith, "The Grand Rapids Dip Revisited," Accident a.n.a.lysis & Prevention, Accident a.n.a.lysis & Prevention, vol. 26, No. 5 (1994), pp. 64754. vol. 26, No. 5 (1994), pp. 64754.

ratio is even higher: Evans, Traffic Safety, Traffic Safety, op. cit., p. 44. op. cit., p. 44.

the rate is .36: David Gerard, Paul S. Fischbeck, Barbara Gengler, and Randy S. Weinberg, "An Interactive Tool to Compare and Communicate Traffic Safety Risks: Traffic STATS," Center for the Study and Improvement of Regulation, Carnegie Mellon University, Transportation Research Board 07-1332, November 2006.

to prove that they are: They also kill others more often. A study in the United Kingdom found, for example, that pedestrians were roughly 1.5 times more likely to die when they were hit by a male driver than a female driver. Car Make and Model: The Risk of Driver Injury and Car Accident Rates in Great Britain: 1994, Car Make and Model: The Risk of Driver Injury and Car Accident Rates in Great Britain: 1994, Transport Statistics Report (London: HMSO, 1995). Transport Statistics Report (London: HMSO, 1995).

more likely to drink: National Inst.i.tute on Alcohol Abuse and Alcoholism. "Drinking in the United States: Main Findings from the 1992 National Longitudinal Alcohol Epidemiologic Survey (NLAES)," U.S. Alcohol Epidemiologic Data Reference Manual, U.S. Alcohol Epidemiologic Data Reference Manual, vol. 6 (Bethesda, Md.: National Inst.i.tute of Health, 1998). vol. 6 (Bethesda, Md.: National Inst.i.tute of Health, 1998).

less likely to wear helmets: C. Peek-Asa and J. F. Kraus, "Alcohol Use, Driver, and Crash Characteristics Among Injured Motorcycle Drivers, Journal of Trauma, Journal of Trauma, vol. 41 (1996), pp. 98993. vol. 41 (1996), pp. 98993.

those who are sober: See, for example, R. D. Foss, D. J. Beirness, and K. Sprattler, "Seat Belt Use Among Drinking Drivers in Minnesota," American Journal of Public Health, American Journal of Public Health, vol. 84, no. 11 (1994), pp. 173237. vol. 84, no. 11 (1994), pp. 173237.

attributed to the driver: Emmanuel Lagarde, Jean-Francois Chastang, Alice Gueguen, Mireille Coeuret-Pellicer, Mireille Chirion, and Sylviane Lafont, "Emotional Stress and Traffic Accidents: The Impact of Separation and Divorce," Epidemiology, Epidemiology, vol. 15, no. 6 (November 2006). vol. 15, no. 6 (November 2006).

and gender differences): G. Whitlock, R. Norton, T. Clark, R. Jackson, and S. MacMahon, "Motor Vehicle Driver Injury and Marital Status: A Cohort Study with Prospective and Retrospective Driver Injuries," Injury Prevention, Injury Prevention, vol. 10 (2004), pp. 3336. vol. 10 (2004), pp. 3336.

Spain to California: See, for example, T. Reuda-Domingo and P. Lardelli-Claret, "The Influence of Pa.s.sengers on the Risk of the Driver Causing a Car Collision in Spain: a.n.a.lysis of Collisions from 1990 to 1999," Accident a.n.a.lysis & Prevention, Accident a.n.a.lysis & Prevention, vol. 36 (2004), pp. 48189, and Judy A. Geyer and David R. Ragland, "Vehicle Occupancy and Crash Risk," UCB-TSC-RR-2004-16, Berkeley, Inst.i.tute of Transportation Studies, 2004; paper accessed at vol. 36 (2004), pp. 48189, and Judy A. Geyer and David R. Ragland, "Vehicle Occupancy and Crash Risk," UCB-TSC-RR-2004-16, Berkeley, Inst.i.tute of Transportation Studies, 2004; paper accessed at http://repositories.cdlib.org/its/tsc/UCB-TSC-RR2004-16.

if there's a pa.s.senger: Actually, if one is is involved in a crash, a pa.s.senger is still a good bet. The added ma.s.s, it has been suggested, could reduce a driver's fatality risk in a frontal collision by 7. 5 percent. See Leonard Evans, "Causal Influence of Car Ma.s.s and Size on Driver Fatality Risk," involved in a crash, a pa.s.senger is still a good bet. The added ma.s.s, it has been suggested, could reduce a driver's fatality risk in a frontal collision by 7. 5 percent. See Leonard Evans, "Causal Influence of Car Ma.s.s and Size on Driver Fatality Risk," American Journal of Public Health, American Journal of Public Health, vol. 91, no. 7 (July 2001), pp. 107681. vol. 91, no. 7 (July 2001), pp. 107681.

pa.s.sengers in the car: Geyer and Ragland, "Vehicle Occupancy," op. cit.

with pa.s.sengers...o...b..ard: Li-Hui Chen, Susan P. Baker, Elisa R. Braver, and Guohua Li, "Carrying Pa.s.sengers as a Risk Factor for Crashes Fatal to 16-and 17-Year-Old Drivers," Journal of the American Medical a.s.sociation, Journal of the American Medical a.s.sociation, vol. 283 (2000), pp. 157882. vol. 283 (2000), pp. 157882.

held for female drivers): B. G. Simons-Morton, N. Lerner, and J. Singer, "The Observed Effects of Teenage Pa.s.sengers on Risky Driving Behavior of Teenage Drivers," Accident a.n.a.lysis & Prevention, Accident a.n.a.lysis & Prevention, vol. 37 (2005), pp. 97382. vol. 37 (2005), pp. 97382.

their male comrades: Ronald Kotulak, "Increase in Women Doctors Changing the Face of Medicine," Jerusalem Post, Jerusalem Post, August 2, 2007. August 2, 2007.

alcohol-related fatal crash: Information on crashes in Montana and New Jersey is drawn from Rajesh Subramanian, "Alcohol-Related Fatalities and Fatality Rates by State, 20042005," DOT HS 810 686, National Highway Traffic Safety Administration, December 2006: available at http://www.nhtsa.dot.gov.

found on rural roads: Growing Traffic in Rural America: Safety, Mobility and Economic Challenges in America's Heartland Growing Traffic in Rural America: Safety, Mobility and Economic Challenges in America's Heartland (Washington, D.C.: Road Information Program, March 2005). (Washington, D.C.: Road Information Program, March 2005).

any other road: ibid.

(nearly 75 percent in 2005): Chicago Tribune, Chicago Tribune, January 12, 2005. January 12, 2005.

or even while driving: Laura K. Barger, Brian E. Cade, Najib F. Aya, et al., "Extended Work Shifts and the Risk of Motor Vehicle Crashes Among Interns," New England Journal of Medicine, New England Journal of Medicine, vol. 352, no. 2 (January 13, 2005). vol. 352, no. 2 (January 13, 2005).

vehicle on the road: This does not have to do entirely with the vehicle, of course. As Charles Kahane of the National Highway Traffic Safety Administration points out, pickup trucks, at least historically, have tended to be driven more often in rural environments and more often by men-two risk-inflating variables. See Charles J. Kahane, "Vehicle Weight, Fatality Risk and Crash Compatibility of Model Year 199199 Pa.s.senger Cars and Light Trucks," National Highway Traffic Safety Administration Report DOT HS 809 662, October 2003.

other kind of vehicle: See, for example, Gerard, Fischbeck, Gengler, and Weinberg, "An Interactive Tool," op. cit.

pickups also impose: Several hundred people per year in the United States are also killed riding in the unprotected cargo beds of pickup trucks. See C. L. Anderson, P. F. Agran, D. G. Winn, and S. Greenland, "Fatalities to Occupants of Cargo Areas of Pickup Trucks," Accident a.n.a.lysis & Prevention, Accident a.n.a.lysis & Prevention, vol. 32, no. 4 (2000), pp. 53340. vol. 32, no. 4 (2000), pp. 53340.

on drivers of other vehicles: See Marc Ross and Tom Wenzel, "The Effects of Vehicle Model and Driver Behavior on Risk," Accident a.n.a.lysis & Prevention, Accident a.n.a.lysis & Prevention, vol. 37 (2005), pp. 47994. vol. 37 (2005), pp. 47994.

more energy in a crash: See Marc Ross, Denna Patel, and Tom Wenzel, "Vehicle Design and the Physics of Traffic Safety," Physics Today, Physics Today, January 2006, pp. 4954. January 2006, pp. 4954.

drivers of smaller cars: Leonard Evans, "Ma.s.s Ratio and Relative Driver Fatality Risk in Two-Vehicle Crashes," Accident a.n.a.lysis & Prevention, Accident a.n.a.lysis & Prevention, vol. 25 (1993), pp. 60916. vol. 25 (1993), pp. 60916.

"was maintained very well": Thanks to Gabriel Bridger for pointing this out. See http://www.iihs.org for results. for results.

in the New Yorker: New Yorker: Malcolm Gladwell, "Big and Bad," Malcolm Gladwell, "Big and Bad," New Yorker, New Yorker, January 12, 2004. January 12, 2004.

Wenzel have pointed out: Tom Wenzel and Marc Ross, "Are SUVs Really Safer Than Cars? An a.n.a.lysis of Risk by Vehicle Type and Model," Lawrence Berkeley National Laboratory Seminar, July 30, 2002, Washington, D.C. Similarly, the Chevrolet Camaro (or Pontiac Firebird) and Chevy Corvette are equally risky to their own drivers, but the Corvette poses less risk to others. The researchers suspect it may be because of the Corvette's fibergla.s.s body and lower profile, both of which might cause less damage to others.

the statistically safest demographic?: Sometimes the statistics confound expectations. Take the Volvo V70 station wagon and the two-door BMW 3 Series. The first car conjures visions of staid Scandinavian safety and innocuous suburban commuting, while the image of the latter is of a small sports car piloted by the typically aggressive "Beemer" driver. Yet according to the Insurance Inst.i.tute for Highway Safety, from 2002 through 2005, the U.S. fatality rate (per million registered vehicle years) for both cars was identical. I have no way of qualifying the difference, and there are a raft of potential statistical problems, but this leads to all kinds of speculation: Did the BMW have better crash protection? Was the safer driving of the Volvo owner offset by inferior handling? Perhaps Volvo wagons carried more pa.s.sengers or logged more miles? Are BMW drivers better drivers? Or was it just a statistical fluke? As Marc Ross remarked to me in an e-mail correspondence, the relatively small number of fatalities in either car means that any variations in how the data is handled can easily throw off the results. There are a host of small factors that can corrupt the data, he explained: "For example, how long was the model in question on the road in the first year. If the model came out early, then the 'exposure' to crashes was relatively long in that first year. If the model came out late, the exposure was short in that first year. A different complication is that models that don't sell so well tend to stay on the dealer's lot, but some of them get registered [for tax reasons] by the dealer but aren't being driven while they stay on the lot." Therefore they have less exposure to traffic risk than might appear.

more than women: See, for example, Pew Research Center, "As the Price of Gas Goes Up, the Nation's Odometer Slows Down," August 8, 2006; available at http://pewresearch.org.

wear seat belts less often: V. Vasudevan and S. Nambisan, Safety Belt Usage Surveys: Final Project Report Safety Belt Usage Surveys: Final Project Report (Las Vegas: Transportation Research Center, University of Nevada, Las Vegas, 2006). (Las Vegas: Transportation Research Center, University of Nevada, Las Vegas, 2006).

trucks without seat belts: Jeremy Diener and Lilliard E. Richardson, "Seat Belt Use Among Rural and Urban Pickup Truck Drivers," Report 42007, Inst.i.tute of Public Policy, University of Missouri, July 2007.