Bad signalling

About a month or two back, a new set of traffic lights came up on the way from home to my daughter’s school. It’s among the least effective set of lights I’ve seen.

For starters, the signal cycles are too long to justify the traffic, so if everyone decides to follow the lights, the traffic will only get worse.

Then, on two sides of the signal (the east and west approach roads) the lights aren’t displayed properly. There are lights on the left side of the road just before the signal, but no corresponding lights at a height across the intersection. So if someone is too close to the intersection they can’t see the lights.

Most importantly there has been no attempt at all so far to enforce the lights. There wasn’t even a cop standing there in the days leading up to the installation of the lights. The intersection was never really that crowded to necessitate a signal (ok I have limited data on this). And one day suddenly the lights went up and it was expected that people will follow.

People, for the most part, haven’t, and that’s a good thing considering the signal cycle. The east and west sides (who don’t have lights on the opposite side of the intersection to reinforce the signal) started the jumping first. The north and south sides joined.

I usually approach it from the south side, and the only times I stop at the signal is if someone in front of me has decided to obey the signal and doesn’t let me past. And I suppose that as people hit the signal multiple times they too make an independent judgment that everyone is jumping the signal, and that’s the optimal strategy.

I see this as a case of failed mechanism design. A set of traffic lights is good if and only if everyone has an incentive to follow it, or is made to follow it. For example, once a set of lights go up, having a cop stationed at the intersection for a few weeks means that most regular commuters across the intersection will get used to the lights, and start following it. And once you have a critical mass of commuters following the signal, the rest simply obey.

As things stand now, I’m happy that nobody really follows these lights – for if people did, it would only slow things down!

These lights, by the way, are at the intersection of 26th main and 39th cross roads in Jayanagar 9th Block.

Traffic signal policy

Is it fair on the part of the city government to direct people’s choices of routes by imposing a suboptimal timing plan on a traffic signal? Or is the government supposed to respond to demand and design the signals looking at the traffic in various directions? Which is supposed to lead which? Which is the hen and which is the egg?

Yesterday I passsed the airport road – victoria road T-junction on the way to office and noticed that the average perceived waiting time on the Victoria road side was significantly higher compared to that of the two branches of the Airport Road meeting there, which was clearly inefficient. Though it is very likely that it has come about because of a generally poor design of the signal, it could also be by design – because the government wants to disincentivize people from using that route.

Given that we are not yet at the libertarian ideal of  “private roads for everyone”, in most municipal regions, the government has the responsibility to build and maintain roads. And given constraints such as the Braess’s Paradox, occasionally it might actually make sense for the government to occasionally direct traffic, rather than leaving it for a free-for-all. I suppose efforts such as converting roads into one-ways are in the same direction.

So given that the government has the monopoly to “give” roads, does it have the right to “take” away roads? If it does, it means that it is effectively trying to control how and where people move. Isn’t that against stuff like freedom of movement? It’s kinda scary.

But if the government doens’t have the right to “take away” roads, what happens to stuff like one way roads, etc.? After all, when you make a road one way, you are imposing a higher cost on certain people (who are willing to brave the heavier traffic in order to move in the “opposite” direction so that the total cost to society at large goes down. Sinilarly, when you put a road divider, you cut off access to certain intersections which would have been very convenient for some people in the larger interest.

So does an elegant solution exist to this problem, assuming that we cannot put tolls on all roads? In principle, putting tolls on roads is fair because we are already taking toll from motorists in non-monetary ways – such as by not maintaining the road, or by subjecting them to too many intersections, or by allowing so many vehicles on the road that reduces the speed of the road. Given all these costs that are imposed on motorists, monetizing them is not tough in principle.

However, politically it is a huge issue and is unlikely to happen for several years to come. In this context, does there exist an elegant solution to traffic management and regulations, that can compensate for inconvenience caused to people in the name of interest of the society-at-large?