On our drives to and from Northern NJ over the Thanksgiving break we encountered two rubbernecking delays. On our way there a vehicle fire already on the side of the road snarled Eastbound I-78 from the Jutland exit (at least that was where we were when we were caught in the backup) all the way to the White House exit. Really, there was nothing to see even though the fire was on our side of the highway. The westbound side of I-78 was snarled even worse, spilling on to I-287 when we got off. The folks there must be wondering what the delay was all about especially since by the time they get to the scene of the incident the emergency vehicles would have already gone. It is also unclear to me that the vehicles on westbound I-78 who slowed to take a look could really see anything.
Unfortunately, once the front vehicles slow down, plus the fact that it was the day before Thanksgiving and volume was starting to build up, there is just no way that traffic could have been eased once the first cars start to gawk. The first cars tend to slow down, well, because they are in front and they see nothing ahead of them so the costs of slowing down for them is small. Unfortunately, they are imposing substantial costs on those futher behind them.
On our way back there was a truck fire - this time, not even on the highway in PA I-83 in a truck stop. There was a lot of black smoke but this was off the highway and the emergency vehicles were on the local roads. The back up stretched from the merge of I-81 and I-83 all the way through almost to Harrisburg.
So what can be done? There is no way to tax rubberneckers especially those in the beginning who essentially cause hours of delays for those at the rear. Some googling led to the following:
1. Installing screens around the scene (doubtful that this would work)
2. An impact analysis; from the conclusion and future research section (emphasis mine):
This study is the first attempt to evaluate the rubbernecking impact of accidents on traffic in the opposite direction based on archived traffic and accident data. ... Barriers are an effective way to reduce the likelihood of rubbernecking in the opposite direction and the delay caused by the rubbernecking ... It is also necessary to investigate the role of human factor on rubbernecking. As indicated in the analysis of this study, motorists in peak period tended to create less rubbernecking than in other periods.
3. An example of negative externality in Chapter 6 of Krugman and Wells (or here).