The Atlantic ran an article in the latest issue, Distracting Miss Daisy (July 2008), in which psychology professor John Staddon critiques the design of the American traffic system, namely the surfeit of street signs. Staddon argues that the mess of signs, cascade of speed limits, and consideration of all conditions do more harm than good. Signs on the roadside distract drivers. Traffic lights suggest that drivers don’t need to pay attention to what other drivers are doing. Maximum speeds are set far below the safe speed for roads in good conditions so drivers don’t adequately assess their speed against road conditions on their own.

Staddon’s article rung a bell because of work by a Dutch engineer reported in Wired back in 2004. Hans Monderman is described in the first sentence as “a traffic engineer who hates traffic signs.” His basic design philosophy is that to make roads safer they have to be percieved as more dangerous. Roundabouts replace traffic signals, sidewalks become level with streets, and extra lanes disappear. The result is that people start paying attention.

For readers not quite following, Steven Landsburg argues in his book The Armchair Economist that the mandated introduction of seat belts actually results in more deaths (driver, passenger, and pedestrian). It has been a long time since I read that passage so I’m not prepared to fully explain or defend that particular argument (based on traffic fatality data), but the line of reasoning familiar to economists is that safety features (physical, financial, or what have you) create moral hazard. People behave in riskier ways when you decrease the immediate risk. Hence some economist at some point argued that we replace airbags with pointed spears as a way of increasing driver attentiveness and decrease overall fatalities.

This behavior pattern follows to many scenarios, and the American roadway is ripe for an overhaul. What makes for an interesting thought experiment, though, is to abstract this idea of too many signs and too many precise instructions and see how it looks elsewhere, beyond traffic systems. The tax code comes to mind. Information systems, especially collaborative information systems, come to mind. The dizzying littany of requirements forced against users causes them to focus on the requirements at the expense of the end goal. The result is meaningless compliance with rote rules.

At root are any number of causes. In the case of traffic signs I think you’d find many constituents unwilling to see warning signs removed even if you could show evidence that the end result is safer. We seem to have a human bias favoring intent over results; the safety net is important even if its placement results in more injury than before. It shows we care. We also seem to suffer from an inability to appreciate “emergent” systems (for lack of a better word - see Eliezer Yudkowsky’s critique of “emergence”). This is why it took so long for the mainstream media to take Wikipedia somewhat seriously and why executives are so slow to adopt open collabortive software in their businesses. It’s why scientific reactionaries reject evolution. It’s why capitalists cry out for government interference (that and naked greed). There’s no way that the process of natural selection could result in a new more “advanced” species. There’s no way that an undirected market could appropriately allocate a scare resource. There’s no way that hundreds of faceless strangers collaborating on a document could produce something meaningful and valid.

Fortunately for us, these “emergent” systems do produce working outcomes. But we reduce their performance when we add too many explicit instructions, and in some critical cases we create moral hazard.


Forbes lists Washington, DC as the most congested metropolitan area in the US. When you compare the size of the DC metro area to that of the other cities on the list, e.g. New York, Houston, Chicago, LA, this is particularly galling.