30 January 2008

The Victory Project

From Arnold Kling:

A web site called The Victory Project proposes billion-dollar prizes

" To the first person(s) that solves any of these Problems:

1. Develop a cure for breast cancer.

2. Develop a cure for diabetes.

3. Reduce greenhouse emissions from petroleum powered automobiles by 95% without increasing the cost of a normal car more than 5%.

4. Achieve 150 miles per gallon of gasoline in a 3,000 lb. car, using EPA standards; without increasing the cost of a normal car more than 10%."

I added my own comment saying that, while I like the basic principle, I think the goals need to be thought through a bit more. They should be broader: rather than reward cures for breast cancer and diabetes they should target something along the lines of increased longevity and wellbeing of, say, the US population, perhaps as measured by Quality Adjusted Life Years. Otherwise the prize might divert scarce research resources away from areas where they can achieve a much better return per dollar. Similarly with 3: total environmental damage is the problem, not just greenhouse gas emissions. Number 4 looks better but, again, the problem is total environmental damage: lots of cheap-to-run vehicles could raise net petrol consumption as well as total environmental damage.

29 January 2008

Catastrophe bonds versus Social Policy Bonds

Catastrophe bonds are typically issued by insurers. The investors who buy them are paid a high rate of interest. If a defined catastrophe (eg a hurricane or pandemic) does not occur, then the investors will make a healthy return on their bond purchase. But if the catastrophe does occur, then the principal initially paid by the investors is forgiven, and is instead used by the insurer to pay its claims to policyholders. (Source) Cat bonds can be quite profitable to investors.

They bear some similarity to Social Policy Bonds. A government could issue Social Policy Bonds that would reward people if, say, a disastrous hurricane did not occur. Holders of the bonds would then be in a similar position to holders of catastrophe bonds: they win if there is no catastrophe. However, Social Policy Bonds would be defined differently. Their objective is to encourage people to do what they can to prevent the catastrophe occurring. So rather than target forces of nature like hurricanes, they might target the numbers of people killed or made homeless by such an event. Unlike holders of cat bonds, investors in Social Policy Bonds would have incentives to reduce the probability and scale of the defined risk.

All is not lost - yet - in Kenya, but the news is grim:
A month after its disputed presidential election, Kenya remains deeply divided and unstable. Politically motivated killings, hackings and gang rapes continue in the towns and in volatile country districts. The economy is faltering. ....What has happened is not a genocide, nor is Kenya anywhere close to being a failed state. But the killings and clearances have been grisly and wretched. When the dead rotting in the maize fields or pulled apart in the wilds by hyenas are eventually counted, over 1,000 Kenyans are likely to have been killed since the election. More than 250,000 have been displaced. The Economist (subscription)
The question that leaps to mind is: why isn't anyone with financial clout interested in insuring against or preventing this sort of catastrophe? The answer's just as clear: there aren't any significant financial investments at stake, as there are when a hurricane strikes the shoreline of a rich country. But that's where the people (or institutions) who could issue Social Policy Bonds come in: they wouldn't know how to avoid a catastrophe like Kenya, but they put up the funds that would motivate investors to do whatever they can to avoid one. The catastrophe bond concept works in practice. So too could Social Policy Bonds, and the potential beneficiaries could include the world's poorest, most vulnerable populations.

27 January 2008

Hunger and aid

From the current 'Economist' (subscription):
[D]ealing with hunger hardly requires a doctorate in the biochemistry of the human body. Breast-feeding advice, food supplements and better hygiene all make a big difference. Most countries know what to do and run pilot programmes that work. But they rarely find the money for full-scale national efforts; the international outfits that might help are ... fragmented and dysfunctional. ... [M]oney for improving nutrition would be the most effective sort of aid around. At the moment, roughly $300m of aid goes to basic nutrition each year, less than $2 for each child below two in the 20 worst affected countries. In contrast, HIV/AIDS, which causes fewer deaths than child malnutrition, received $2.2 billion—$67 per person with HIV in all countries (including rich ones).
On what basis are aid funds allocated? Availability of tv footage? The caprice of celebrity donors? Political correctness? All probably play a part. The one criterion that doesn't seem relevant is efficiency. Of course, it's not quite that simple: many of the obstacles to rational resource allocation are probably third world governments that have little interest in looking after their populations, and every interest in syphoning off aid funds or otherwise obstructing, for their own narrow purposes, aid workers. That probably diverts resources away from some of the areas in most desperate need. A Social Policy Bond regime targeting basic health indicators in the developing world could go a long way toward redressing these perverse incentives, perhaps by persuading recalcitrant governments to take long golfing holidays.

Meanwhile it's worth comparing the £300 million that the rich world gives to alleviate hunger in the poor countries to the amount it lavishes on its own farmers: in 2006 that amounted to $268 billion according to the OECD. Yes, you read that right: billion. To put it another way: taxpayers in the rich world give 89 times as much to their own agriculture sector as they do to starving people in the third world. In return we get: a devasted rural environment, the destruction of wildlife and bloated oligopolistic agribusiness corporates, and thousands of highly skilled lobbyists adept at maintaining the status quo.

24 January 2008

Exporting environmental damage

I've often wondered how much of the undeniable success of western economies depends on exporting our problems to the rest of the world. I've known about the effects of our insane agricultural support policies on developing countries for some time: the closing our our markets; the catastrophic effects of exports of subsidised overproduction on farmers; the export of price instability for farm products - see this (pdf) for an extreme statement. Now here is a report about our environmental impact on the third world:
The environmental damage caused by rich nations disproportionately impacts poor nations and costs them more than their combined foreign debt, according to a first-ever global accounting of the dollar costs of countries' ecological footprints. The study, led by former University of California, Berkeley, research fellow Thara Srinivasan, assessed the impacts of agricultural intensification and expansion, deforestation, overfishing, loss of mangrove swamps and forests, ozone depletion and climate change during a 40-year period, from 1961 to 2000. In the case of climate change and ozone depletion, the researchers also estimated the impacts that may be felt through the end of this century. "At least to some extent, the rich nations have developed at the expense of the poor and, in effect, there is a debt to the poor...."
What this implies for the development path of the poor countries should be obvious. Unfotunately policymakers in those countries, for the most part, persist in targeting aggregated economic goals - principally GDP and growth of GDP - which, apart have little to do with the wellbeing of ordinary people, threaten to replicate the environmental depredations of the west. Far better would be to target meaningful human (as distinct from economic or corporate) goals, such as those incorporated in the Human Development Index: life expectancy, standard of living, education, literacy. We need to remind ourselves of what we already know: economic growth is not an end in itself.

22 January 2008

Education is not schooling

Short excerpts cannot do justice to Jay Griffiths' Wild: An Elemental Journey, but this is a blog about policymaking, rather than a book review. Here (page 140) she is writing about the Inuit in Canada :
School is not a synonym for education. Your might, if you're lucky, get a bit of an education at school, but for Inuit children, the land was their education. White lawmakers forced Inuit children to go to school, insisting that their parents settle in communities.... One result is that people are dependent on store-bought food, and if they have no cash they go hungry. ... a stark physical example of the effects of not knowing the land. But the psychological effects are everywhere. Without knowledge, you cannot be out on the land. without survival skills, you can barely set foot beyond the perimeter of the community. Young people are effectively imprisoned by this ignorance into the small and claustrophobic communities where they go stir-crazy.
The policymaking mentality hasn't changed much: government isn't content merely to raise and allocate funds for the (laudable) goal of educating children. It has a single, astonishingly limited vision as to what the form and substance of that education shall take. Naturally it's biased in favour of its own educational experience: the sort that leads to careers in lawmaking. It's the same in other fields: government views crime as something to be tackled by the police and justice system. Health is something for a Ministry of Health to deal with. Mental health is about more psychiatrists, counselling and drugs. The policies are all neat and compartmentalised, just like the bureaucracies and the state of mind that generates them. The real world, though, is too messy - wild - for that. Effective policies need to stimulate people's imagination; to cope with society's as different from our own as the Inuit; and to deal with rapidly changing knowledge and circumstances. Broad goals are stable over time: the most effective ways of achieving them are not.

Social Policy Bonds would allow governments to do what they are good at: articulating society's goals and raising the revenue to achieve them. But the actual achievement of these goals would be far more efficiently done by the market. Currently the markets' formidable incentives serve the private goals of shareholders and business operators. A Social Policy Bond regime, in contrast, would channel the market's efficiencies into the service of public goals.

20 January 2008

Irrationality as a policy driver

US Presidential candidate Barack Obama speaks:
We have been operating under a politics of fear: fear of terrorists, fear of immigrants, fear of people of different religious beliefs, fears of gays that they might get married and that somehow that would affect us. We have to break that fever of fear … Unfortunately what I've been seeing from the Republican debates is that they are going to perpetuate this fearmongering …It's absolutely true there are 30,000, 40,000 hard-core jihadists who would be happy to strap on a bomb right now, walk in here and blow us all up. You can't negotiate with those folks. All we can do is capture them, kill them, imprison them. And that is one of my pre-eminent jobs as president of the United States. Newsweek
Not all fears are irrational of course, and of course there are some things we ought to fear more than we do and make policy accordingly (species extinction springs to (my) mind). All the same, as Sharon Begley continues:
The fact that a candidate whose campaign is built on optimism and a positive message is not above evoking terrifying images of suicide bombers and nuclear bombs—and doing so two breaths after he denounces fearmongering—reveals the power of fear to sway voters. Half a century of research has shown that fear is one of the most politically powerful emotions a candidate can tap, especially when the fears have a basis in reality; jihadists, of course, are indeed bent on suicide bombings. ... "In politics, the emotions that really sway voters are hate, hope and fear or anxiety," says political psychologist Drew Westen of Emory University, author of the recent book The Political Brain: The Role of Emotion in Deciding the Fate of the Nation. "But the skillful use of fear is unmatched in leading to enthusiasm for one candidate and causing voters to turn away from another."
Fear can easily over-ride our rationality. For instance, we fear plane crashes more than we ought, and don't fear car crashes enough, and unfortunately policymakers react accordingly. Social Policy Bonds, which would target broad outcomes - accidental deaths, in this instance - would lead to more rational policy than that likely to follow a spectacular plane crash - especially one for which copious visual footage is available.

17 January 2008

Procedure is king

More from Bad Food Britain:
'Hazard Analysis Critical Control Point' (HACCP) ... focuses on identifying the 'critical points' in a process where food safety problems - hazards - could arise... What HACCP boils down to is a system of checklists, form filling and record keeping. ... This system creates a paper trail so that in the event of a problem, the companies or producers implicated can demonstrate that they did their bit and walk away blameless, plausibly denying responsibility. ... A supermarket that poisoned customers by selling contaminated chicken... could use HACCP to show that its suppliers followed correct procedures so it was not at fault.
The overriding objective - the one that permeates the entire system - has little to do with food safety and everything to do with protecting oneself from disciplinary proceedings or lawsuits. There's a similar confusion in other policy areas. Here is Bruce Schneier on airport security:
Surprising nobody, a new study concludes that airport security isn't helping: "A team at the Harvard School of Public Health could not find any studies showing whether the time-consuming process of X-raying carry-on luggage prevents hijackings or attacks. They also found no evidence to suggest that making passengers take off their shoes and confiscating small items prevented any incidents."

And: "The researchers said it would be interesting to apply medical standards to airport security. Screening programs for illnesses like cancer are usually not broadly instituted unless they have been shown to work."

Note the defense by the TSA: "'Even without clear evidence of the accuracy of testing, the Transportation Security Administration defended its measures by reporting that more than 13 million prohibited items were intercepted in one year,' the researchers added. "Most of these illegal items were lighters.'"

This is where the TSA has it completely backwards. The goal isn't to confiscate prohibited items. The goal is to prevent terrorism on airplanes. When the TSA confiscates millions of lighters from innocent people, that's a security failure. The TSA is reacting to non-threats. The TSA is reacting to false alarms. Now you can argue that this level of failures is necessary to make people safer, but it's certainly not evidence that people *are* safer.
Again, rational policy is subordinated to procedure. It can't work: we need diverse approaches that can adapt to changing circumstances. Such pluralism are anathema to centralized policymaking. But what government policymakers can do is specify target objectives, and contract out their achievement to the market. Actions should serve outcomes - not procedures. A Social Policy Bond regime would ensure that all actions would aimed at achieving the social and environmental goals set for it by government. Being freely marketable, they would encourage the adaptive, diverse solutions that big, complex social and environmental problems demand.

16 January 2008

Too many doctors?

What would be the typical government reaction to a health care crisis? More doctors, of course. Shannon Brownlee questions this, at least for the US. Unlike the number of car dealers, for instance,
which depends on the number of people who want cars and can afford them ...how much health care patients want or need has far less influence over the supply of physicians. That’s because for the most part it’s your doctor and not you, the consumer, who determines how much care you receive. When your doctor says you need a CT scan, you get one. When your doctor says you should go to the hospital, you go. Doctors, in effect, generate some of the demand for their services, so that even when there are large numbers of them per capita, they can keep their appointment books full. Source
There's also the possibility that fewer doctors could lead to better care:
More tests and procedures always entail more risk, and for care that’s unnecessary, the ratio of benefit to risk is zero. What’s more, where numerous doctors, particularly specialists, are routinely involved in a patient’s case, the potential for miscommunication and confusion multiplies. Modern medicine should be a team sport, but it is often practiced as if everybody is running a different play. Different doctors order duplicative tests, prescribe drugs that interact poorly with what the patient is already taking, and assume another physician will attend to a critical aspect of a patient’s care. A cardiologist can be a virtuoso at slipping a stent into the coronary artery of a patient in the throes of a heart attack, but if she leaves it to another physician to prescribe aspirin to her patient -- one of the most effective treatments for preventing a second heart attack -- that prescription might fall through the cracks.

This is what appears to be happening in many hospitals, where the ratio of specialists to primary-care physicians is especially high. In one recent study, two Harvard economists ... examined how the quality of care in different states varied as the proportion of specialists rose. They found that measures of quality, like the percentage of heart-attack patients who received a prescription for aspirin, tended to fall in direct proportion to a rising ratio of specialists. The point ... “is not that the specialist is inferior, but that the system is not accounting for the ‘coordination cost’ specialists are imposing.”
This is worrying because the likely policy response is going to create more such problems. How could a politician, even one convinced by solid research proving Ms Brownlee's points, cut back training programmes for doctors? How would it look to his or her political opponents, or to the media, when a medical mistake occurs at a hospital that would have had more doctors but for our politician's cutbacks? It's not just health care. In every policy area the penalties for doing something likely to succeed but different far outweigh the benefits for doing something that has been tried before and shown to fail. Rising crime? Spend more on police or CCTV cameras. Traffic congestion? Build more roads. Not enough fish being landed? Subsidize more powerful fishing boats. The incentives are to placate lobby groups and vested interests, not to achieve society's goals.

15 January 2008

Policy and rationality

From Bad Food Britain, by Joanna Blythman:
Rarely a week goes by on British television screens without a programme detailing, in disturbing detail, the unsavoury underbelly of industrial food production, or the effect that cheap, overprocessed junk food has on the nation's health. A minority ...is so affected by that knowledge that it changes its shopping habits instantly, boycotting this or that, and forking out more for an alternative. But within a few days after the headlines die down, it is business as usual for most British shoppers. page 133
So too in the world of policy. Attention spans are short, and what we don't really want to know, we can easily bury under the blizzard of new information. Slow-moving, unglamorous and inconvenient facts - and I don't just mean about climate change - are demoted in our minds. Politicians react to crises or events for which there is compelling tv footage. They are not always the ones to which we'd give highest priority if we were being rational. One solution would be to specify policy in terms of the outcomes it's supposed to achieve. Such outcomes will be less responsive to ephemeral events or distractions or the goals of corporate bodies or other lobby groups. On the same theme, that of rationality in policy, here is another idea, and US citizens could follow it up. You could sign this petition to be sent to Congress:
I am not afraid of terrorism, and I want you to stop being afraid on my behalf. Please start scaling back the official government war on terror. Please replace it with a smaller, more focused anti-terrorist police effort in keeping with the rule of law. Please stop overreacting. I understand that it will not be possible to stop all terrorist acts. I am not afraid.

14 January 2008

Anecdote is not proof

It's almost a given that government spending is inefficient. But it is not sufficient to show high absolute and relative government expenditures coexist with social and environmental problems, as they do in the rich countries. The facts are, though, suggestive.

  • Government spending in the rich countries typically amounts to at least a third of national income.
  • Many social indicators have shown negligible improvements in recent years. Life expectancy, infant mortality, hospital outputs, literacy, violent crime, suicide, poverty and income inequality have barely changed despite a massive increase in social spending.
  • Around the world there is little relationship between higher public spending and better social outcomes.

At first sight, the conclusion that government is inefficient seems unarguable, but in fact there are counter-arguments: sure the indicators listed above haven’t improved much, but there are others that have. Or: there are mitigating circumstances, such oil price rises, larger numbers of people on welfare, aging populations, etc. So while there is widespread perception that we are not getting value for money from the taxpayer dollar, there’s little in the way of proof.

More convincing than anecdote, to me anyway, is the persistence of perverse subsidies: those policies, like agricultural support programmes, that transfer funds from the poor to the rich, that environmental depredation, and that are financially disastrous. They were set up with good intentions, but they have helped finance the creation of lobby groups whose raison d'etre is resistance to their withdrawal. Their persistence, in the face of decades of evidence as to their failure, points to a more dangerous truth: that mechanisms that correct disastrous programmes can fail. Indeed, one result of poor policymaking is that it widens the gap between policymakers and ordinary people - a problem in itself and one that is likely to generate more problems further down the road.

10 January 2008

Unnecessary surgery

An article in Men's Health about unnecessary surgery in the US shows how far we are from outcome-based health policy:
[In November 2006] at the American Heart Association's annual scientific sessions in Chicago, a tremor rumbled through through the McCormick Place convention center. Turns out the cause of the quake (which only cardiologists could feel) was a major new study on angioplasty, that near-miracle procedure in which a balloon-tipped catheter is used to unblock an artery. The seismic finding? Only that angioplasty is no more effective than medication for a large segment of heart-attack victims. Or, to put it in stark statistical terms, heart doctors have been performing as many as 50,000 unnecessary operations every year. Frightening, yes, but nowhere near as scary as the broader implications. When the air was suddenly let out of the angioplasty balloon, there was at least one cardiologist present in the convention center who recognized the researchers' conclusion as just another symptom of a condition he'd seen before: doctors cutting first and asking questions later. Is Your Surgeon Scamming You?, John Brant
It's disappointing that cash incentives trump ethics in such circumstances, but we could at least have policies that bias the incentives in favour of long-term positive health outcomes, rather than a profusion of invasive, often-dangerous, procedures that do nothing to benefit the average patient. Unfortunately governments tend to focus on meaningless targets whose only virtue is that they are easy to measure, such as throughput in emergency wards, ambulance response times etc. Targets like these are easily gamed, and while they are quantities that can be measured and compared, they are costly to gather and have nothing to do with the welfare of patents. Of course there would be difficulties in subordinating health policy to meaningful measures of wellbeing (such as Quality Adjusted Life Years) but if carefully designed it would I think lead to a far more successful healthcare system.

08 January 2008

Corporate goals are not societal goals

Susan Pagan’s 9-year-old daughter recently made the honor roll, but when the Florida mom saw the report card, she was appalled. There on the envelope was a cartoon of Ronald McDonald along with a potential “food prize” for elementary school students who had good grades, behavior or attendance.

“Reward yourself with a Happy Meal!” the report card jacket urged. And to further associate fast food with praise, approval and success in the minds of young consumers, the offer also stipulated that the “report card must be presented at the time of ordering.” Julie Deardoff 16 December 2007
This sort of thing happens when we assume that corporate objectives are necessarily congruent with society's. They aren't. Neither does a strong economy, as measured by such flawed indicators as Gross Domestic Product, mean a successful society. Unfortunately, in the absence of clear, broad goals for the education - and nutrition - of our children, the corporations feel they can move in. And they're not mistaken:
The saddest part about this whole story? [Susan] Pagan was told that she was the only parent who thought it was inappropriate to put fast-food ads on the report card jackets and that the district would consider her complaint next year.

05 January 2008

Avoiding catastrophic climate change

How do you put a value on extreme events when doing cost-benefit calculations? Extreme environmental events will become more common if the climate changes radically. They might have a low probability of occurring, but they could be catastrophic. Jim Giles in the New Scientist says that economists have generally ignored extreme events when doing cost-benefit calcuations about climate change: they are so unlikely and lie so far in the future that it is not cost-effective to try to prevent them. "But environmental groups argue that the risk of extreme events justifies large investment now...."
Martin Weitzman of Harvard University has developed the first thorough method for including unlikely but extreme events, such as widespread crop failures, in cost-benefit analyses. When you take into account extreme temperature rises of more than around 6 °C, he says, they dominate all other options and effectively demand that investment aimed at stopping them be made now. "This tells us that we should take the problem much more seriously that normal cost-benefit analyses suggest," says Weitzman.... It "probably means we should spend more money now, but it doesn't tell us how much."
Exactly. The scientific uncertainties are huge and so is the range of plausible expenditures. So too is the urgency of the problem and any policy debate under the current system is quite simply going to take too long. The usual international initiatives demand a degree of scientific and economic certainty that is higher than we are ever likely to achieve until the problem has become too late to do much about. The current system is fundamentally flawed: it insists on too high a degree of proof before action can be taken. It does so because our governments are the ones that bear all the risk of making a wrong decision. Governments have to find the (considerable) resources to spend now on a problem that may or may not materialize years or decades in the future. Naturally they will be reluctant to do so until the evidence becomes impossible to ignore - by which time it will be too late to avert a calamity.

Climate Stability Bonds would solve that problem by transferring the risks of getting it wrong to those willing to bear it - voluntarily and with a continuing incentive to get the cost-benefit calcuations correct. Globally-backed Climate Stability Bonds would contract out the achievement of climate stability to the most efficient operators. If they fail to perform, then it is they who would lose out, not taxpayers the world over. It would be in their interests to do the calculations correctly, and to continuously refine them in line with our rapidly expanding knowledge and technology. They would have incentives to spend resources immediately on the most effective climate-stabilising projects, while bearing all the risk of failure. They certainly wouldn't delay doing anything until catastrophe becomes imminent and obvious to recalcitrant taxpayers...by which time it's also likely to be inevitable.

04 January 2008

The unimportance of being responsible

John Kay writes:
In politics, business and finance, as on the seas, the hero is the person who tackles a problem, rather than the person whose actions prevent the problem arising. ...If Margaret Thatcher had acted to deter Argentina from invading the Falklands, rather than ordering a taskforce to remove the occupying forces after they had landed, she would probably have been remembered as an unsuccessful one-term prime minister.
He's right: being remembered and winning elections are not as helpful as preventing problems arising in the first place but unfortunately they constitute success in our current policymaking system. As a society we'd do better to reward those of us who anticipate and avoid social and environmental problems before they become emergencies. A Social Policy Bond regime that rewarded the maintenance of the best aspects of the status quo could do this: bonds could target, for example, the absence of large-scale wars, or use of nuclear weapons; or the absence of catastrophic climatic events or large-scale disease epidemics.

Many of our social and environmental problems need long-term, unglamorous, patient, adaptive and diverse approaches to their solution. Such approaches seldom cover their practioners with glory or even recognition - still less do they win elections. There are many well-meaning people and organizations in these areas and many of them do superb, heroic jobs with few resources. A Social Policy Bond regime could both enlarge this pool of effective problem-anticipators and divert more resources their way. Incentives do matter and it would be a good idea, I think, if the people currently devising ingenious advertising campaigns for dogfood were instead given the chance to provide decently for their children by working to deter nuclear warfare or mass environmental disasters.

01 January 2008

We're all Aztecs now

Cars have done much to destroy life, not only directly but also indirectly by taking over our cities, poisoning our air, destroying communities... and the list goes on. Here's yet another angle, from the New York Times via Grist:
Motorized outdoor enthusiasts are converging in increasing numbers on Western public lands [in the United States] - not only in areas marked for such outdoor enthusiasm, but in wilderness areas where rules against off-roading are nearly impossible to enforce. Registration of all-terrain vehicles and motorbikes in four Western [US] states tripled from 1998 to 2006. The surge is traceable to the booming outdoor-recreation industry, as well as the culture of sprawl: In some places, houses have been pushed out so far that federally owned land is just a big backyard - albeit a public backyard where no individual has to take the specific blame for vehicle-aggravated erosion and water pollution. Off-roaders deny criticism that they're out to defile untouched nature, arguing that public land is there for public use. "[Groups lobbying for wilderness designations] think it has to be kept in this pristine state," says one motorcycle-shop owner. "These people don't even use it." Which is, of course, the point.
Regarding the header of this post:
Michael Harner, in his 1997 article The Enigma of Aztec Sacrifice, estimates the number of persons sacrificed in central Mexico in the 15th century as high as 250,000 per year. Source
That's towards the higher end of the range of estimates. Worldwide, about 1.2 million are killed on the roads annually.