Image for post
Image for post

But does it change behaviour?

Some interventions are so obvious that they don’t need justifying. Or do they?

Your correspondent only drives intermittently, and then mostly on motorways. That does not make for the most riveting journeys, but one advantage of highway driving is that there are so very few road signs. In case of roadworks, you encounter speed limit and narrow road signs, but other than that it is the visual equivalent of peace and quiet for miles and miles. One would almost forget about the existence of some rarer signs — like the one I encountered on a trip through rural Devon last week.

As I left the motorway to join the A361, I noticed a wild animals crossing sign. This road is, at that point, a fast dual carriageway, with a speed limit of 70mph. It winds through the countryside with, presumably, a variety of roaming wildlife occasionally wishing to traverse the road. In the US, I once saw the damage a collision with a deer can do to a pick-up truck, so it seemed perfectly sensible to warn drivers of such a dramatic eventuality.

And then I started to wonder. The principal purpose of traffic signs is to influence road users’ behaviour, but I couldn’t immediately pinpoint how exactly this sign would influence mine. Imagine a deer unexpectedly jumps out onto the road. At a speed of 70mph, I would have driven at least 20m before I can begin to brake. From then it would take me around 6 seconds and another 70–80m to come to a standstill. It’s hard to see how I could avoid colliding with anything that suddenly appeared less than 50m in front of me.

Image for post
Image for post
Could you stop in time if it jumped on the road? (photo: Virginia DOT CC BY)

The only thing I could do to significantly reduce the risk of an accident would be to slow down — a lot. But then, would it not be better to simply put in place a 30mph speed limit? Looked at it in this way, the wild animals sign seemed rather pointless — much like its sibling, the falling rocks sign, which raises the same question: what are drivers supposed to do with the knowledge that rocks might fall? Both cases even carry a risk of backfiring: nervous drivers could take their eyes off the road to look for signs of wildlife ready to jump from the thicket, or of boulders about to fall from the rock face.

An apparently reasonable measure turns out at best futile, and potentially even more hazardous than simply doing nothing. And this is not a unique situation.

In Belgium, my native country, the debate around drink-driving has flared up, after the brand-new speaker of the Flemish parliament had to resign, having been caught driving with three times the legal maximum level of alcohol in his blood (0.05%). Calls for zero tolerance (no alcohol whatsoever) were loud and widespread, and quite understandably so. In Flanders, one in three road fatalities is the result of accidents with alcohol as a material cause, and so in principle all of these deaths could be avoided if road users completely stayed off the booze.

Some people object to such a measure, often with poor arguments: zero tolerance leaves other causes of road deaths (texting while driving, driving while tired) untouched, and people might inadvertently breach the zero limit if they had a liqueur chocolate or mussels in white wine. There are also some poor arguments in favour, for example, the idea that any amount of alcohol impairs one’s driving ability, with meaningless statements like “with 0.05% alcohol in your blood you have a 40% higher chance of having an accident than when sober” (apart from being alarmist, it tells us nothing if we don’t know what this chance actually is when sober). Such arguments don’t help settle the question whether the measure would be effective. Is there any evidence that a tougher limit works (or not)?

The World Health Organization regularly publishes a comprehensive report on road safety around the world, which may shed some light on the matter. Looking at a selection of countries that are more or less comparable to Belgium, we can assess whether a lower alcohol limit corresponds with fewer alcohol-related road deaths.

Image for post
Image for post
Alcohol-related fatalities as a proportion of all road traffic fatalities (source: WHO)

The picture is far from clear. The left-hand figure shows the proportion of alcohol-related road fatalities, with countries shown in bands according to their legal maximum limit. Ireland and Belgium are the bad apples in the 0.05% group, doing worse even than the countries where the limit is 0.08%. In the countries with a lower limit, the proportion of alcohol-related fatalities is lower, but with the intriguing exception of China, they do not significantly perform better than the best countries with a higher 0.05% limit. There is a weak correlation (R 2=0.14 — perfect correlation would be 1) between the legal drink-drive limit and the proportion of alcohol-related fatalities

The report also includes a rating for the effectiveness of enforcement (on a 0–10 scale, indicated in brackets for each country). The right-hand figure lists countries according to this measure. Ireland is an extreme outlier where, despite a rating of 10, excessive alcohol was involved in nearly 40% of all road fatalities. If this country is excluded, the correlation between enforcement and the percentage of alcohol-related road fatalities is slightly higher than with the legal limit (R 2=0.16), but still very modest.

Image for post
Image for post
Tougher limit, or tougher enforcement? (source: WHO)

The effect of enforcement does seem to be more pronounced than that of reducing the limit. Moving from a rating of 6 to one of 8 would reduce the proportion of alcohol fatalities by 1/3, and cranking enforcement up to 10 halves it. Reducing the limit from 0.05% to 0.02% would lead to a reduction in alcohol fatalities of just 20%.

Nevertheless, the wide variation between countries, even with similar limits and enforcement, suggests that drink-driving is a complex affair, in which these two instruments do not tell the whole story: the correlations are too weak to predict with any certainty that either intervention would fix Belgium’s serious drink-driving problem.

But would either of the two measures affect driver behaviour?

The target group in this case are drivers who violate the current limit. Cutting this limit to 0.02% or even 0% is a bit like reducing the speed limit in a village from 30mph to 20mph, to stop people driving at 50mph. Would a stricter limit make speed merchants slow down or boozers drink less? It’s not so easy to see how that would work.

Image for post
Image for post
Maybe the right-hand glass keeps me below the 0.02% limit? (photo: Maarten Takens CC BY)

The simplicity of the message that not drinking alcohol is the only safe choice is appealing, but it is unlikely that drivers who breach the limit now do so because the 0.05% limit is too complicated. Moreover, a lower limit would require a much larger group — those whose consumption is within the current limit — to change their behaviour too. Will they rapidly throw out the old social norms where 1 glass of wine or beer with the meal is fine? That is not so certain. There is also the risk of the slippery slope: if one drink gets you over the new limit anyway, then “in for a penny, in for a pound”, you might as well have a few more.

Especially if the chance of being caught is perceived to be low. Perhaps the root of Belgium’s drink-driving problem lies there, rather than in not having zero tolerance. Its enforcement rating of 6 out of 10 is unimpressive. Increasing the perceived likelihood of being checked (and caught out if over the limit) is a more plausible behavioural mechanism than lowering the legal limit. The WHO data do not unequivocally demonstrate this, but they do support it.

Just imagine that you knew there was a 1 in 10 chance that a deer would jump out of the bushes to cross the road whenever you drove down the A361. Would you reduce your speed way down from 70mph — or still take a chance?

If we want to change our own, or other people’s conduct, we should resist unthinkingly going for the obvious, self-evident interventions. We should ask how a proposed measure will actually influence decisions and behaviour. If that is not clear, we should keep looking, rather than pursuing potentially counterproductive window-dressing.

Originally published at on July 19, 2019.

Thanks for reading this post — I hope you enjoyed it. Please do share it far and wide — there are handy Twitter and Facebook buttons nearby, and you can click here to share it via LinkedIn, or simply copy and paste this link. See all my other posts (I publish one every week) here. Thank you!

Written by

Accidental behavioural economist in search of wisdom. Uses insights from (behavioural) economics in organization development. On Twitter as @koenfucius

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store