The Tullock Spike — or Why Killing You May Make You Safer

Counterintuitive, Episode 1

I’ve finally had the time to enter podcasting! If you’re up to it, join me for journeys into stories that are not what they seem to be. Episodes examine unusual events from a broad spectrum that will surprise you and then make you think. Because there’s always a layer beneath. You can find new episodes on Spotify (and basically any other major podcasting platform) or SoundCloud. Below, you’ll find the transcript of the first episode with some reference / further reading hyperlinks.

— — — — — — — — — — — — — — — — — — — — — — — —

September 3rd, 1967. Stockholm, Sweden. It’s a quiet day. Too quiet. Especially considering that, in the preceding weeks and months, all hell was expected to break loose on September 3rd. From experts to ordinary citizens, predictions for September 3rd were dire. TV crews are on stand-by, ready to capture anything going wrong at a moment’s notice. They call this Dagen H, or H-Day, in Sweden. A huge project, many months in the making and clocking in at around 300 million euros in today’s money, they call this the most important Swedish day in 1967.

It was the day when Sweden switched driving lanes, from left to right.

From the viewpoint of the Swedish government, surrounding European countries now predominantly used the right lane, more and more cars were imported into Sweden from countries driving in the right lane and accidents on the border to right lane-driving countries were ever-increasing. Change needed to happen. Imagine, an entire country, from one day to the next, changes something as fundamental as on which side of the road cars are driving. Imagine driving in the left lane your entire life and then changing directions in the span of 24 hours, doing everything the ‘wrong’ way. No wonder experts predicted a veritable bloodbath with a historic spike in accidents.

So what happened on September 3rd, 1967 in Sweden? Well…not much. Despite grave expectations, the whole operation was a resounding success with no fatalities. Even more astounding, the year 1967 in Sweden had fewer fatal traffic accidents than the years before or after. Don’t be mistaken, this is not an isolated phenomenon. One year after this, Iceland switched driving lanes on a single day as well. The casualties of this dramatic change? One single boy on a bicycle, who fell and broke his leg.

To understand this counterintuitive outcome, we need to leave the Nordics for a moment and travel to Ontario, Canada. Here, Dutch émigré and professor of psychology Gerald Wilde proposed his theory of risk compensation. Let’s take driving. In a study, researchers equipped a fleet of taxis in Munich, Germany, with anti-lock brakes, ABS. They monitored this group of taxis as well as another group, which didn’t receive ABS, for 36 months. The ABS system is certainly superior with regard to safety. It decreases stopping distance and prevents skidding. Intuitively, you would therefore expect the superior braking system to result in a lower accident rate. So what did the researchers discover? The opposite. The rates of accidents were nearly identical between the two groups, with a slight lead of the group with the ABS.

How can a braking system which is objectively better not result in a lower accident rate? Enter Gerald Wilde. In 1982, Wilde proposed that people adjust their risk behaviour according to their perceived level of safety. Keep that ‘perceived’ in mind, it will come in handy later on. If they notice something becoming safer, they start to up the ante on riskiness. Because there is no immediate short-term benefit from more safety, while there frequently is a reward for acting more risky. So the Munich taxi drivers newly upgraded with the safer ABS system decided to drive more aggressively and break later than they would have with their old equipment. The reward: getting to their destination more quickly and earning more money.

There is a bit of a controversy about risk compensation which is rooted in a slight misunderstanding. Some people interpret risk compensation as negating each and every advance in safety technology by a proportionate increase in risky behaviour. Yet even though this was the outcome of the Munich taxi driver study, it’s not the case for most safety measures. Usually, estimates put the effect of risk compensation at around 40–50% of the total effect. So, on average, the increase in risky behaviour nearly halves the expected safety benefit. But that still means that, given all the safety improvements in cars in the last decades, the risk of death while driving has considerably decreased, just at a slower pace than in a world without risk compensation.

Even more interesting, this risk-seeking reaction to increased safety only holds true in situations with a perceived increase in safety. This could open doors to clever design which hides increases in safety from the user for more effectiveness by circumventing risk compensation. But it could also lead to a compounding of risk in case of faulty airbags for instance, where you get the disadvantages of risk compensation without reaping the benefits of superior safety technology.

Risk compensation is pervasive and there is hardly an area in which it doesn’t pop up, if you pay attention. Skiers wearing helmets are more risk-prone than those without, yet of course helmets decrease the risk of head injury in case of a fall. An increase in perceived safety leads to an increase in risky behaviour but an overall improvement in safety. American football players tackle other players with their helmet-protected heads in a dangerous manoeuvre known as spearing, incurring the very same head injuries which the helmet was designed to prevent. Spearing was officially banned in 1976 but, and here I quote from a position statement of the National Athletic Trainers’ Association, ‘helmet-contact rules are rarely enforced and head-down contact still occurs frequently’. Because it yields a competitive benefit in football, spearing and other head-endangering techniques are used by players. Winning first, safety second. Think about how eager you would be to ram someone dressed in hard safety gear with your unprotected head versus with an armoured shell around your head.

Risk compensation falls into the broad category of unintended consequences and is one of the many red herrings that broadcast our limited understanding of the incredibly complex world around us. But these unintended consequences don’t have to be bad. There is no universal law stipulating bad consequences in all circumstances. The trick is to identify the silver lining of our mistakes, understand them and make use of them for public policy or our personal lives.

So, what’s the silver lining of risk compensation? Imagine, for a moment, that we install a giant metal spike on top of your steering wheel so that, every time you crash into something, the spike will impale and kill you. Would that make you behave more or less risky as a driver? For most people, the prospect of imminent death makes for an excellent reason to be cautious in their behaviour. This thought experiment was created by professor of law and economics Gordon Tullock at George Mason University. The eponymous Tullock spike gives us our silver lining. In the face of a perceived decrease in safety, people decrease the riskiness of their acts. Risk compensation. Again, this doesn’t have to be a real decrease in safety for the mechanism to work. Go ahead and tell people the airbags of a rental car are defunct. Regardless of whether they do work or not, people will likely drive a lot more cautious in response to that.

With this in hand, we can now return to Sweden and Iceland. When driving lanes were switched on September 3rd, 1967, the level of perceived uncertainty must have hit record levels. Going through every motion in the ‘wrong’ direction must feel extremely stressful and unsettling. Even worse, there’s no room for a break. You have to stay alert all the time, always consciously checking whether it’s the right direction. It’s not only you though. There are also the other drivers which you have to pay attention to, in case they make a mistake. This cocktail of vigilance and worry is perfectly suited for avoiding accidents on the road. You perceive a decrease in safety, so you avoid risky behaviour. Risk compensation.

Basically, this is the timeline in Sweden. 1966: People drive in the left lane and are used to it. 1967: Driving is switched to the right lane, people are unsettled, more cautious, and cause fewer fatal accidents. 1968: People drive in the right lane and are used to it. So much so in fact, that road traffic deaths in 1968 nearly reach the levels of 1966 again. Because, inherently, there is not much of a difference between driving in the left or right lane; it’s mostly convention. The major difference lies in the perception of the situation. In 1967, driving in the right lane is perceived as unfamiliar and dangerous. In 1968, it’s the new norm. The level of risky behaviour is no longer actively decreased, the level of attention paid to surroundings and oneself no longer heightened. Accident rates resume their ascent to previous heights.

If the decrease in risky behaviour upon less perceived safety is our silver lining, can we actively make use of it? Can we design public spaces to capitalize upon risk compensation, to create literal safe spaces? As it happens, we can. A design concept known as shared space does exactly that. Remove the demarcation between cars and pedestrians, remove as many traffic signs as you can get away with, create the feeling of uncertainty. To anyone unfamiliar with risk compensation, this would seem crazy! ‘Irresponsible’ would probably be the nicest thing people would say to you if you would propose such a design. But you know better now. You know that such a public design aims to tune into risk compensation and will most likely result in a lowered accident rate. At least for a while. But even if residents of a city which implemented shared spaces might get used to this uncertain situation, tourists and anyone passing through the city would still find themselves subject to risk compensation, yielding long-lasting value of shared space.

This doesn’t have to be restricted to urban planning either. Remember the football players with their concussion problems? The naïve solution of just removing their helmets might stop the headbutting and tackling, but it would also expose their heads and potentially lead to serious head injuries after the unavoidable crashes in American football. But we can do better than that. How about this: just remove the helmet from players during training sessions but allow them to wear their helmets during games. Training sessions present a more controlled environment than the rough-and-tumble setting of a game. Even more important, players learn their moves during training. So if they are used to having a vulnerable head during training, they won’t engage in headbutting opponents. If they didn’t learn to use this as a common strategy, they also won’t use it in the actual game as they’re not experienced in doing so. You get the best of both worlds: Players stop doing risky head-first tackles and still benefit from the safety advantages of wearing a helmet. Researchers did just that, in a randomized controlled clinical trial, and witnessed a 30 percent decrease in head impacts during games in comparison to the control group which didn’t experience training without helmets. Unintended consequences stemming from mechanisms such as risk compensation don’t have to be unmitigatedly bad. If we just make the effort to understand what’s going on, we can make use of it, in public design, sports and maybe even in your personal life. Who knows, maybe a bit of discomfort would be good for all us, to avoid falling into the trap of convenience and daily routine.

Here’s an interesting bonus which offered a valuable lesson to me while preparing this episode. Initially, I wanted to connect the lower rates of fatal traffic accidents in situations exhibiting decreased perceived safety with the notion of disfluency in psychological research. Proposed in 2007, disfluency was intended to activate the conscious processing of information to avoid intuitive errors. There is this famous bat-and-ball problem in which study participants are asked this: A baseball bat costs $1 more than a ball and together they cost $1.10. How much does the ball cost? Most people go for 10 cents but then the bat would cost $1 dollar more than that, so $1.10, which would clock in at $1.20. Of course the answer to that is that the ball costs 5 cents and the bat $1.05. Don’t worry if you’ve got that wrong, many people do.

In fact, so many that psychologists are looking for ways in which you can force people to really think about the problem before blurting out the obvious but wrong answer. Disfluency was proposed to make a problem slightly harder and thereby force you to engage with it on a more rational level. For the bat-and-ball problem, researchers posed the question in a hard-to-read font in 2011 and reported higher levels of accurate answering compared to an easy-to-read font. The reasoning goes like this: if you have to really concentrate to decipher this font, your brain is already in high-activity mode and can be immediately used for problem-solving. Sounds great, doesn’t it? So great in fact that it caused quite some buzz, including numerous citations and a New York Times article. So great that I really wanted to incorporate it and discuss its implications for risk compensation. So I went on a hunt for recent developments in this field…and found that not only couldn’t this study be reliably replicated but in 2016 there was a meta-analysis debunking the whole phenomenon.

Meta-analyses are the aggregated analysis of all eligible studies published about a certain topic in the last years to decades. Ideally, they are more reliable than individual studies and are used to determine the magnitude, relevance and, sometimes, the existence of an effect. So at least these subtle forms of disfluency (such as using a hard-to-read font) don’t seem to have an effect. More extreme forms of disfluency, such as deleting whole words from sentences, were shown to be more robust in their effect and, if they hold true, could be a mechanism which also appears in risk compensation. Removing traffic signs, analogous to removing words from a sentence, could lead to more active processing of the situation and thereby reduce the amount of careless errors made while driving. Science is a continuous process and if I would record the same episode in a couple of years, I probably would need to change it quite a bit. But if you ask me, that’s the beauty of it.

I hope you’ve enjoyed this first instalment of Counterintuitive! If you did, join me next time, where we’ll talk about communication and the nature of power. You can find references and further reading for this episode in the show notes. If you like Counterintuitive, please recommend it to your friends and give it a 5-star rating on Spotify, Stitcher or wherever you get your podcast from. It really helps. A new episode will be uploaded every two weeks. My name is Daniel Bojar and you’ve listened to Counterintuitive, the critical thinking podcast about things which are not what they seem to be. You can follow me on Twitter at @daniel_bojar or on my website dbojar.com, where you will find articles about more counterintuitive phenomena. Until next time!