Thursday 31 October 2013

It gets worse

Last weekend, I noted Nick Wilson's suggestion that we ban the import of cars that can't block cellphone use inside the car.

His letter to the NZ Med Journal, on which his blog post was based, has a few other interesting bits. He has a few other policy suggestions for reducing cell phone use by drivers. Here they are:
  • All new mobile phones permitted on the New Zealand market from 2018, could be required to automatically disable themselves from working when their internal GPS sensor identifies movement (albeit with an exemption for phoning the national emergency number). This option could work alongside the “smart car” option above, or may obviate its need. It could potentially help prevent injuries among people who use electronic devices while cycling (for whom injury risks appear elevated10).
  • Introduce new regulations that increase fines and/or other penalties for infringements of the existing law. International evidence has shown this to have a strong deterrent effect and is key to maintaining the effectiveness of laws prohibiting drivers’ use of mobile phones. One option includes mobile phone confiscation from those using them while driving.
  • Address the residual need to prohibit hands-free phones in cars given the incontrovertible evidence, collected internationally7 and in New Zealand,11 that these are also highly distracting for drivers. ... Exemptions for commercial drivers and emergency workers could still be permitted, once drivers' [sic] demonstrate appropriate knowledge around hazard mitigation (e.g. how to keep to short sentences when conversing).
So commercial drivers who can pass some kind of test showing they can speak in short sentences could use cell phones.

I agree with Wilson that there's little difference between hands-free and regular cell phone use; some of us argued when this first came out that it was a slippery slope towards banning all in-car phone use.

But the worst part is this:
Nevertheless, given that passing a law is not particularly expensive in New Zealand (e.g. typically at NZ$3.5 million12,13; 95%CI: 2.0–6.2 million), it would not take long for such a new law to be cheaper than one or two mass media campaigns. But the new law would probably also be much more cost-effective than media campaigns if its effects lasted many decades into the future.
Cost-effectiveness cannot be measured solely against costs borne by the government. The mass-media campaign would cost money for the ad in addition to the "costs of passing a law" figure Wilson cites. But a law banning the import of non-compliant cars or non-compliant phones would impose very large costs on everyone who uses phones in cars. And a technology that blocked passengers in cars or buses from using their phones would impose very large costs on those passengers. Those costs matter too.

Rather a lot of public health advocacy can only be understood if we start from the perspective that the costs of regulation consist only of costs on the government and that regulation does not impose any cost on the regulated.

In comments over at SciBlogs, Wilson notes that there were, in 2012, 59 deaths and injuries associated with mobile phone use in New Zealand cars. Let's use that number to see what possible benefit Wilson could here be pursuing.

While the 2012 crash stats aren't yet out, there's no reason to expect that proportions changed since 2011. In 2011, there were 284 deaths, 2060 serious injuries, and 10514 minor injuries resulting from car accidents. If we apportion those proportions to Nick Wilson's 59 deaths and injuries, and round up in Wilson's favour, we get 2 deaths, 11 serious injuries, and 45 minor injuries. NZTA estimates the social costs of fatalities at $3.8 million, serious injuries at $650k, and minor injuries at $64k.

Let's further assume that each and every one of these was an innocent bystander outside of the driver's car. I'd be happy to argue that we shouldn't count costs drivers impose on themselves and that passengers are in a Coasean bargain with the driver, but then Nick would accuse me of leaning too heavily on rationality and we'd get into a fight about that rather than a fight about his rather silly policy proposals. So let's say every one of these is an external cost: there was never a phone-using driver who was harmed through his own silliness and all costs were instead imposed on others.

The total social costs are then $17.6 million. Let's round that up to $20 million to allow for any incidental costs. Let's further assume that every one of us bears some of that $20 million in total risk because we don't know which innocent will be killed or injured by a phone-using driver. Over 4.4 million people, let's say that rounds to $5 per capita.*

I consequently expect that the absolute largest possible benefit from any of Nick's proposals around phone use in cars is $5 per capita per year. I'm assuming that every crash he identifies was 100% caused by the phone and wouldn't have happened anyway. I'm assuming that his policies would be 100% effective and that every one of those crashes disappeared. I cannot make a better case in his favour. And it's $5 per capita per year. I'd not be surprised if the true benefit were less than half of that figure. But let's stick with $5 per capita per year.

Wilson noted one cost of his proposal: legislation costs about $3.5 million to pass. I'd argue that this confuses average with marginal and that the correct opportunity cost of legislation is rather the legislation displaced by any piece of legislation. But in any case, we'll take it as a one-off that won't continue.

The bigger cost is the cost his policy would impose on each and every commuting passenger who would prefer to be using a cell phone. ANZA notes that there are 2.5 million cell phones subscribed to data plans in New Zealand. So 2.5 million people are potentially affected. Let's round that down and say half the population. Suppose further that only half of them would ever use a mobile device in a car or on a bus. If each of them gets at least $20 in value from using the phone while in a car, Wilson's proposal destroys value even before we start factoring in any costs of policy implementation, enforcement, costs imposed on people buying new compliant cars, costs imposed on people buying new compliant phones, or costs imposed on people who have to go out and get a dedicated GPS because their Google Maps no longer work because Nick Wilson thought he had a good idea.

All of that in pursuit of a $5 per capita benefit.

Wilson, in comments, writes:
Nevertheless, other non-technological solutions might ultimately prove to be more cost-effective in reducing mobile phone use while driving (eg, better enforcement and higher fines). What’s probably needed is a thoughtful societal discussion about the options and the pros and cons of each option. We should not be too quick to rule out any particular option – given the complexities and trade-offs that need to be made between convenience, cost and preventable deaths.
No. No no no no no. Policies that seem likely to fail cost-benefit analysis by orders of magnitude should not be considered. They should be ruled out immediately. At best, public discussion wastes everyone's time and energy. At worst, something destructive gets implemented. Should Pharmac give serious consideration to homeopathy?

* Anticipating another potential critique of the use of these cost-of-life measures, I will simply note that all of these figures are Ministry of Transport standard. They use them in deciding whether straightening a curve or adding a passing lane is worth the cost, balancing reduced crash costs against road construction costs. If we abandon them for one aspect of road safety, we throw the whole darned system out of whack. I'd not be surprised if Nick's proposals imposed costs of at least $50 per capita once we accounted for what it does to cell phone or car markets. If we spend $200 million to save 2 lives when we could instead have used the $200 million in total cost to fix a traffic black spot that kills 5 people, then we've really really screwed things up even if you hate the idea of putting dollar values on lives.

10 comments:

  1. One additional word on the disable-while-moving issue: "Buses"

    ReplyDelete
  2. This poor guy seems a bit hung up on mobile phones, is it really the most pressing public health issue he could put his mind to?

    ReplyDelete
  3. Ayup. Thought I'd included them among "each and every commuting passenger", but more than happy for it to be explicit.

    ReplyDelete
  4. He also wants to redistribute income to reduce health inequalities, abolish tobacco, add a pile more regs on drinking, fight global warming as a public health initiative. If you just modeled it as "Public health excuses to let me be the boss of other people", you'd replicate a lot of his policy proposal outputs. Expect that wouldn't be how he'd see his objective function, but a machine running that algorithm would provide pretty similar output.

    ReplyDelete
  5. Don't be too hard on the NSA, Eric, I hear agent HGW XX/7 is a swell guy.

    ReplyDelete
  6. There was a time when 'public health' meant something done by one person that affected someone else. e.g. a person who walked around with an infectious disease. These folk at https://blogs.otago.ac.nz/pubhealthexpert/ have expanded it to things like 'eating' and smoking which can only affect the eater/smoker. Bad driving can be a hazard to others so we would like to have fewer bad drivers - but not sure that using a cell phone at a red light amounts to a hazard to others.

    ReplyDelete
  7. Yeah, I was surprised by this turn when I first encountered it back around 2006 - it was from before then, but I'd not noticed. Basic premise now that the main burden now comes from "non-communicable" diseases so that's where their focus ought to go.

    It's hardly just Otago though - a lot of public health researchers around the world spend a lot of time on tobacco control, alcohol control, obesity regulation/taxation, and so on. For many, it's all wrapped up with various social justice concerns.

    See here for ex http://econlog.econlib.org/archives/2006/07/physician_consu.html

    ReplyDelete
  8. Thanks for the references

    ReplyDelete
  9. Ahh jeez what is wrong with you people. Here in Thailand it is an acknowledged fact that the people can easily ride a motor cycle with four or five on board, carry some goods and fodder in one hand, and talk on the cell phone, all while driving on the wrong side or the middle of the road. We know this, the minister of Science told us that Thais were gifted this way. He also told us that there would be no floods in Thailand this year.

    ReplyDelete
  10. I agree that this guy is really in the 'boss other people about' camp when it comes to policy - if it moves, pass a law against it.


    But back to my earlier point on your weekend post... innovation will probably solve a lot of this due to the integration coming between cars and 'devices'... Horace Deidu at Asymco.com has a big theory about integrating mobile technology and cars.


    If we follow Nick Wilson's suggestions then NZ would effectively outlaw innovations that could solve the problem without imposing costs on anyone...


    As it is I use my time parked on Auckland's most centrally located car park (motorway) each morning to catch up on podcasts... now under Nick Wilson's model would that be banned too?

    ReplyDelete