ISET

ISET Economist Blog

A blog about economics in the South Caucasus.

An Economist’s Comment on “Dodge or Die” on the Streets of Tbilisi

[“Dodge or Die” is a series of reports on the relationship between the pedestrian and the motorist in Georgia by Robert Linkous].

As Stephen Dowling put it in his BBC News article a few years ago, “when it comes to crossing the road, there's no such thing as an international standard. Every country does it differently.” How people drive and cross the road, according to Dowling, is a matter of a country’s cultural values. Is it really?

Bad traffic habits are a grave problem in many cities around the world. Yet, while many have managed to largely solve the problem, in Georgia, the problem remains virulent and it is not even fully appreciated. Observing the traffic dynamics in a couple of main avenues, like Rustaveli or Chavchavadze, is sufficient for a first time visitor to Tbilisi to realize that something is wrong with the Tbilisi traffic system. Yes, that’s right – something is wrong with the system, not culture!

Game Theory, a field of applied mathematics that is used to analyze “strategic interactions” between agents who pursue conflicting goals, can possibly explain what is going on in the streets of Tbilisi. Strategic interaction is just a fancy name for what happens in games, such as chess, or in real life situations such as bargaining over the price of a taxi ride. Luckily, Game Theory also points to a possible policy solution.

The first relevant insight from Game Theory is that the behavior of traffic participants is determined by their expectations about the behavior of other actors. For instance, if pedestrians don’t expect drivers to stop at a zebra crossing, they will hesitate to cross until they are 120% sure that the car approaching them is actually stopping. Given the very steep price for making a judgment mistake, the result is a very stable behavioral pattern. Pedestrians won’t cross. As someone who makes every effort not to do as the Romans do in Rome, I rediscover this time and again whenever I am slowing down before a zebra crossing. It takes intensive hand waving and actual stopping for the pedestrians to understand my intentions and adjust their expectations accordingly.

The same logic applies to the behavior of drivers. Any driver familiar with the Tbilisi rules of the game would not slow down before a zebra crossing because it makes no sense. Pedestrians would in any case not trust his/her intentions and would patiently wait for the car to pass (or stop, in the rare case it is driven by a foreigner like myself). Moreover, it makes sense for a Tbilisi driver to speed up whenever he or she observes a group of pedestrians getting ready to cross the road on a zebra. Why? Because by speeding up s/he sends a clear signal about his/her behavior – hey, I am not going to stop, don’t even try crossing!

The second key insight from Game Theory is that this kind of “uncivilized” outcome (drivers not giving way and pedestrians not even trying to cross on a zebra) is stable. In other words, it does not make sense for drivers and pedestrians to change their “strategy” (behavior) given what they know about the other party’s strategy. The implication is that Tbilisi will be forever stuck with this uncivilized “equilibrium” (another fancy term used by game theorists). Unless someone (the policymaker) bothers to change the rules of the game and reset expectations. More on this later.

By now it should be easy to see the possibility – observed in most European and North American cities – of a situation that is the exact mirror image of the situation in Tbilisi: drivers respecting pedestrians’ rights and pedestrians being quite assertive about exercising them. Not only is this civilized equilibrium possible, it is also quite stable. If drivers expect the pedestrians to cross on a zebra, they will start slowing down well before the first pedestrian sets his/her foot on the road. Expecting this type of behavior, pedestrians will not hesitate to cross. The outcome is stable because it makes no sense for German drivers to start behaving Tbilisi style (they will quickly find themselves behind bars) and there is no point for the German pedestrians to hesitate before crossing.

Two crucial questions are these:

1) How come some countries are stuck in a bad kind of equilibrium while others are able to enjoy the benefits of civilization?

2) Can a country (or city) permanently shift from one type of equilibrium to another?

Before I proceed with a formal “solution”, let me say that the same exactly questions apply to many other areas of strategic interaction among people. For instance – and this is an important hint – it applies to petty corruption, e.g. the interaction among bribe givers and takers. And as we know from Georgia’s recent experience (both good and bad), a determined policy intervention or the breakdown of law and order can swiftly shift a country from one equilibrium to another. Moreover, a determined policy action can lead to a permanent adjustment of expectations, and, yes, a change in culture!

The obvious policy solution is to introduce tougher regulations, higher fines and stricter enforcement concerning both drivers and pedestrians. As far as traffic regulations are concerned, Georgia has already had a positive experience with the introduction of fines for not wearing seatbelts. The policy worked extremely well and led to an instantaneous change in the drivers’ behavior. Ideally, strict rules should be enforced for all kinds of violations including the widespread practice of parking on sidewalks – forcing pedestrians to share their already limited space with cars. An effort should be made to improve the underpasses they are often dirty and insufficiently illuminated, giving incentives to jaywalking. The fines for jaywalking should be increased from their current, ridiculously low level of 3 GEL.

Understandably, enforcement would have to be particularly strict during a relatively short transition period to allow all traffic participants to properly reset their expectations and behavior. Once expectations are reset, however, the intensity of enforcement (and related costs) could go down quite dramatically because the new equilibrium will be able to sustain itself. At least according to insights from Game Theory.

Rate this blog entry:
31 Comments

Related Posts

Comments

 
Guest - Sanjit Dhami on Monday, 10 June 2013 15:51

Nice post and a very pressing problem indeed. My views below are not meant in any way to be criticisms of this article but rather as trivial clarifications of some concepts.
1. The article says: "The first relevant insight from Game Theory is that the behavior of traffic participants is determined by their expectations about the behavior of other actors."
This is not necessarily an insight from game theory. The effect of expectations on one's behavior has always been a part of economics- even classical economics going back of Adam Smith and even formalized in the work of Cournot and others. Even In decision theory, one chooses among lotteries conditional on the expectation that one has about how nature will behave (embodied in the lottery itself).
2. You are right that game theory allows for multiple equilibria. Again this idea was well known from development economics for a very long time- for instance, the idea that some countries may be struck in a poverty trap while others could be prosperous. Indeed precisely on the most important question- if we have multiple equilibria, which one will be chosen- is where game theory fails to provide a satisfactory, empirically confirmed answer (despite many papers written on it). The idea of engineering moves between equilibria is a fascinating one (we dabbled in this issue here: http://www.le.ac.uk/economics/sd106/Philanthropy.pdf).
3. The fact that whenever we observe people not cooperating with each other, we try to invoke a Nash equilibrium outcome and a prisoner's dilemma frame of mind, limits ones intuition and imagination about the problem.

Nice post and a very pressing problem indeed. My views below are not meant in any way to be criticisms of this article but rather as trivial clarifications of some concepts. 1. The article says: "The first relevant insight from Game Theory is that the behavior of traffic participants is determined by their expectations about the behavior of other actors." This is not necessarily an insight from game theory. The effect of expectations on one's behavior has always been a part of economics- even classical economics going back of Adam Smith and even formalized in the work of Cournot and others. Even In decision theory, one chooses among lotteries conditional on the expectation that one has about how nature will behave (embodied in the lottery itself). 2. You are right that game theory allows for multiple equilibria. Again this idea was well known from development economics for a very long time- for instance, the idea that some countries may be struck in a poverty trap while others could be prosperous. Indeed precisely on the most important question- if we have multiple equilibria, which one will be chosen- is where game theory fails to provide a satisfactory, empirically confirmed answer (despite many papers written on it). The idea of engineering moves between equilibria is a fascinating one (we dabbled in this issue here: http://www.le.ac.uk/economics/sd106/Philanthropy.pdf). 3. The fact that whenever we observe people not cooperating with each other, we try to invoke a Nash equilibrium outcome and a prisoner's dilemma frame of mind, limits ones intuition and imagination about the problem.
Guest - Y on Monday, 10 June 2013 15:57

"Understandably, enforcement would have to be particularly strict during a relatively short transition period to allow all traffic participants to properly reset their expectations and behavior. Once expectations are reset, however, the intensity of enforcement (and related costs) could go down quite dramatically because the new equilibrium will be able to sustain itself."
Very true, but...unless we live in autocracy or a "managed democracy", which ruling party will put their reelection at risk by actually enforcing these rules? Tough measures require spending considerable political capital (as Georgia learned from the experience of reining in corruption and crime). Without overwhelming public support for these measures, the regulators will see the game's simply not worth the candle.

"Understandably, enforcement would have to be particularly strict during a relatively short transition period to allow all traffic participants to properly reset their expectations and behavior. Once expectations are reset, however, the intensity of enforcement (and related costs) could go down quite dramatically because the new equilibrium will be able to sustain itself." Very true, but...unless we live in autocracy or a "managed democracy", which ruling party will put their reelection at risk by actually enforcing these rules? Tough measures require spending considerable political capital (as Georgia learned from the experience of reining in corruption and crime). Without overwhelming public support for these measures, the regulators will see the game's simply not worth the candle.
Guest - Eric Livny on Monday, 10 June 2013 18:40

A valid concern, but I don't think that the amount of political capital to be spent here is all that large. And there is some political capital to be gained as well, particularly if tough measures are implemented soon after the election.

A valid concern, but I don't think that the amount of political capital to be spent here is all that large. And there is some political capital to be gained as well, particularly if tough measures are implemented soon after the election.
Guest - Giorgi on Monday, 10 June 2013 22:00

Thing is that whoever enforces or should enforce these new rules is mostly a driver himself. So there is no utility gain for him in enforcing anything of the kind - precisely because he thinks in the terms given by Levan below. So easy as it may be on paper/in theory, I highly doubt that it can be easily done in practice.

Thing is that whoever enforces or should enforce these new rules is mostly a driver himself. So there is no utility gain for him in enforcing anything of the kind - precisely because he thinks in the terms given by Levan below. So easy as it may be on paper/in theory, I highly doubt that it can be easily done in practice.
Guest - Giorgi on Monday, 10 June 2013 22:07

Another important point is that you can sell anti-bribery measures easily, but 90% of driver citizens of Georgia will never give up the god-given right of running the red light or not stopping at zebra crossings :))

Another important point is that you can sell anti-bribery measures easily, but 90% of driver citizens of Georgia will never give up the god-given right of running the red light or not stopping at zebra crossings :))
Guest - Hans Gutbrod on Friday, 12 December 2014 20:54

Eric, I bet that a large ISET class could change the behavious in Tbilisi, if they work on it for 3 months and create a vivid and salient campaign.

It would be a great example of policy from below, to fill a political gap. Students would learn huge amounts about social behaviour and changing "rules of the game". (I might even volunteer to teach that class...)

Eric, I bet that a large ISET class could change the behavious in Tbilisi, if they work on it for 3 months and create a vivid and salient campaign. It would be a great example of policy from below, to fill a political gap. Students would learn huge amounts about social behaviour and changing "rules of the game". (I might even volunteer to teach that class...)
Guest - Giorgi Kelbakiani on Tuesday, 11 June 2013 15:58

This problem (and such problems in other areas) cuts very deep actually. Maybe I am going to be too philosophic here but...
"which ruling party will put their reelection at risk by actually enforcing these rules?"
that is THE question. And I would say my answer on it is "every second (or third)".
If not dealt, such problems start to overwhelm the society increasing the feeling "this country needs a strong and strict hand" among people rising a need for radical changes and dictatorial power (and radical changes and dictatorial power are the things that are often achieved by a revolution).
Hence, the government that is not willing to risk to enforce these rules is risking to give a green light to another dictatorial regime, which will be willing to risk, but will likely do most things in a wrong fashion (I won't list the reasons why)... and then end up like most dictatorships do, being overthrown (again often by revolution).
So there is a big chance getting trapped in a circle. The government, which is likely to break that circle, is the one that manages not to loose connection to the society, by justifying very well all those strict regulations and what is important, delegating some responsibilities down to the people by raising awareness that many things are fault of society and not just government. There is no other way, I think.
I do not know how many revolutions Georgia will need to realize this and to achieve the well functioning democracy where everyone takes their share of responsibility, I only know that France required five (though it was a couple centuries ago).

This problem (and such problems in other areas) cuts very deep actually. Maybe I am going to be too philosophic here but... "which ruling party will put their reelection at risk by actually enforcing these rules?" that is THE question. And I would say my answer on it is "every second (or third)". If not dealt, such problems start to overwhelm the society increasing the feeling "this country needs a strong and strict hand" among people rising a need for radical changes and dictatorial power (and radical changes and dictatorial power are the things that are often achieved by a revolution). Hence, the government that is not willing to risk to enforce these rules is risking to give a green light to another dictatorial regime, which will be willing to risk, but will likely do most things in a wrong fashion (I won't list the reasons why)... and then end up like most dictatorships do, being overthrown (again often by revolution). So there is a big chance getting trapped in a circle. The government, which is likely to break that circle, is the one that manages not to loose connection to the society, by justifying very well all those strict regulations and what is important, delegating some responsibilities down to the people by raising awareness that many things are fault of society and not just government. There is no other way, I think. I do not know how many revolutions Georgia will need to realize this and to achieve the well functioning democracy where everyone takes their share of responsibility, I only know that France required five (though it was a couple centuries ago).
Guest - Eric Livny on Tuesday, 11 June 2013 16:47

:)

[The latest French republic was established fairly recently (not through a revolution though)].

:) [The latest French republic was established fairly recently (not through a revolution though)].
Guest - Giorgi Bakradze on Tuesday, 11 June 2013 17:43

France had a new republic every time they lost a war and were saved by the English, so they don't really count :)))

France had a new republic every time they lost a war and were saved by the English, so they don't really count :)))
Guest - Giorgi Kelbakiani on Wednesday, 12 June 2013 15:13

What you mean by "saved by the English", at least couple times they restored monarchy there, not republic.
Anyway, certainly the problems french had back then and the problems we have now might be :) different (no surprise there). Nevertheless, it is always problems (both small and big) that are not dealt, what leads to angry society who seeks guilt in government. That is what I wanted to say basically.

What you mean by "saved by the English", at least couple times they restored monarchy there, not republic. Anyway, certainly the problems french had back then and the problems we have now might be :) different (no surprise there). Nevertheless, it is always problems (both small and big) that are not dealt, what leads to angry society who seeks guilt in government. That is what I wanted to say basically.
Guest - Levan on Monday, 10 June 2013 16:22

Very interesting analysis. The broader question that emerges from this is "civilized" vs "uncivilized". All uncivilized behavior might be very logical and practical from the point of view of behaving agent, until one introduces moral values. So, for example, giving cars the right of way would seem only logical and practical, if we did not consider moral values:

- The cost of mistake is higher for pedestrians, so they may want to be in control.

- The financial cost of avoiding a mistake (stopping) is higher for a car in terms of wear of brakes and tires, energy wasted on stopping and accelerating back a vehicle that weighs a ton. Pedestrians do not need 70 meters of stopping distance and 12 seconds.

- Cost in time lost: To let 10 cars pass, a pedestrian would only need to wait for 3 seconds. To let 10 pedestrians pass, a car would "waste" about ten times longer: 5 seconds to decelerate, 20 seconds to wait, and 10 seconds to accelerate.

So, why do people introduce morality to mess up perfectly logical equilibria? Does morality have a benefit, or has it no explanation at all?

Very interesting analysis. The broader question that emerges from this is "civilized" vs "uncivilized". All uncivilized behavior might be very logical and practical from the point of view of behaving agent, until one introduces moral values. So, for example, giving cars the right of way would seem only logical and practical, if we did not consider moral values: - The cost of mistake is higher for pedestrians, so they may want to be in control. - The financial cost of avoiding a mistake (stopping) is higher for a car in terms of wear of brakes and tires, energy wasted on stopping and accelerating back a vehicle that weighs a ton. Pedestrians do not need 70 meters of stopping distance and 12 seconds. - Cost in time lost: To let 10 cars pass, a pedestrian would only need to wait for 3 seconds. To let 10 pedestrians pass, a car would "waste" about ten times longer: 5 seconds to decelerate, 20 seconds to wait, and 10 seconds to accelerate. So, why do people introduce morality to mess up perfectly logical equilibria? Does morality have a benefit, or has it no explanation at all?
Guest - Eric Livny on Monday, 10 June 2013 19:59

Levan, I have children ages 7 and 8. I cannot let them walk around the neighborhood (Vera), as simple as that. The cost of mistakes is indeed higher for them (and me) than for the occasional Tbilisi driver. Unfortunately, they may naively attempt to use the zebra crossing for its intended purpose and get killed. How does children's life enter your practical/financial equation?

Levan, I have children ages 7 and 8. I cannot let them walk around the neighborhood (Vera), as simple as that. The cost of mistakes is indeed higher for them (and me) than for the occasional Tbilisi driver. Unfortunately, they may naively attempt to use the zebra crossing for its intended purpose and get killed. How does children's life enter your practical/financial equation?
Guest - Giorgi on Monday, 10 June 2013 21:52

You forget to count in a cost of years in prison if the driver actually runs a pedestrian over. And no, drivers shouldn't have the right of way, they enjoy the advantage of, well, being drivers.

You forget to count in a cost of years in prison if the driver actually runs a pedestrian over. And no, drivers shouldn't have the right of way, they enjoy the advantage of, well, being drivers.
Guest - liatodua on Monday, 10 June 2013 19:46

Installing button-triggered traffic lights would be more effective. And cheaper if you take the admin.costs into consideration.

Installing button-triggered traffic lights would be more effective. And cheaper if you take the admin.costs into consideration.
Guest - Dace on Monday, 10 June 2013 22:07

Very nice written, Eric! I have been thinking along the same lines. Just some comments from the practical side of changing the culture (from the experience of Latvia). Latvia did indeed manage to switch to the good equilibrium with stricter fines and enforcement. However, there is more work for economists in calculating the optimal amount of zebra crossings on a busy road. For a poor country, the cheapest way is just to paint some white lines on a road. Accelerating and slowing down every 50 metres will slow down the traffic and annoy drivers (and here maybe there is some difference between annoyed Latvians and annoyed Georgians). To avoid traffic jams, you will need to somehow reduce the traffic. In Georgia, it can be painful - need to reestablish the technical check of cars and regulate the taxi business - sometimes it seems that 4/5 of all the traffic are taxis, and limit parking (increase fee).

In Tbilisi, there are actually many new air bridges made, only a couple tunnels stink, I also know button-triggered traffic lights... but pedestrians do not use them. Yes, 3 lari fine is too low, but if that was really enforced, I think the situation would improve. I have seen it enforced only once so far and probably only because the policemen saw a beautiful girl jaywalking.

And what about the so popular but useless speed bumps? Small streets of Vera and yards - it's full of bumps but still cars are accelerating between them and it's a country where they love children. They put bumps on highways when they are doing some road works. They put bumps also between crossroads. I think these bumps are clearly a bad equilibrium. Just road signs usually work fine in other countries. Here, there are usually no signs at all - many bad equilibria are created because rules are not conveyed in a clear manner and enforced consistently.

Very nice written, Eric! I have been thinking along the same lines. Just some comments from the practical side of changing the culture (from the experience of Latvia). Latvia did indeed manage to switch to the good equilibrium with stricter fines and enforcement. However, there is more work for economists in calculating the optimal amount of zebra crossings on a busy road. For a poor country, the cheapest way is just to paint some white lines on a road. Accelerating and slowing down every 50 metres will slow down the traffic and annoy drivers (and here maybe there is some difference between annoyed Latvians and annoyed Georgians). To avoid traffic jams, you will need to somehow reduce the traffic. In Georgia, it can be painful - need to reestablish the technical check of cars and regulate the taxi business - sometimes it seems that 4/5 of all the traffic are taxis, and limit parking (increase fee). In Tbilisi, there are actually many new air bridges made, only a couple tunnels stink, I also know button-triggered traffic lights... but pedestrians do not use them. Yes, 3 lari fine is too low, but if that was really enforced, I think the situation would improve. I have seen it enforced only once so far and probably only because the policemen saw a beautiful girl jaywalking. And what about the so popular but useless speed bumps? Small streets of Vera and yards - it's full of bumps but still cars are accelerating between them and it's a country where they love children. They put bumps on highways when they are doing some road works. They put bumps also between crossroads. I think these bumps are clearly a bad equilibrium. Just road signs usually work fine in other countries. Here, there are usually no signs at all - many bad equilibria are created because rules are not conveyed in a clear manner and enforced consistently.
Guest - Eric Livny on Tuesday, 11 June 2013 00:18

Could not agree more, Dace: planning traffic - button-triggered and regular lights, bumps, under and over-passes, zebra crossing, etc., - is a science (it is actually far more rigorous than our beloved economics). There are even formal degree programs that teach the relevant skills. Unfortunately for Georgia, this science is yet to leave its mark on Tbilisi roads. As I drive I notice many thanks that could be improved and wish there was an address for proposals for change. I hope someone in Tbilisi government is reading this blog...

Could not agree more, Dace: planning traffic - button-triggered and regular lights, bumps, under and over-passes, zebra crossing, etc., - is a science (it is actually far more rigorous than our beloved economics). There are even formal degree programs that teach the relevant skills. Unfortunately for Georgia, this science is yet to leave its mark on Tbilisi roads. As I drive I notice many thanks that could be improved and wish there was an address for proposals for change. I hope someone in Tbilisi government is reading this blog...
Guest - Admin on Tuesday, 11 June 2013 00:02

test

Guest - Jeffrey Zilbermann on Tuesday, 11 June 2013 00:04

Pedestrians are moving targets and cars actually speed up when you cross the street. Road culture actually shows the true values of the country, sorry to say! I am sure that there is not more road rage in response to the behavior of drivers who can't drive.

Pedestrians are moving targets and cars actually speed up when you cross the street. Road culture actually shows the true values of the country, sorry to say! I am sure that there is not more road rage in response to the behavior of drivers who can't drive.
Guest - Eric Livny on Tuesday, 11 June 2013 20:24

We are all targets, some moving slow, some fast!

We are all targets, some moving slow, some fast!
Guest - Admin on Tuesday, 11 June 2013 17:06

test

Already Registered? Login Here
Register
Guest
Wednesday, 11 December 2019

Captcha Image

Our Partners