(Reposted from Complexia)
In Addis Ababa there are three main forms of public transport: government-run buses with set routes, contract taxis, and an informal network of buses and minivan taxis that operate on a range of different routes across the city based largely around commuter demand.
I have been a regular user of the minivan taxis for my daily commute over the last seven months, and I have come to view this network as exhibiting many of the characteristics of a complex-adaptive system.
As explained in the first post on Complexia:
A complex-adaptive system is made up of autonomous yet interrelated agents (system components) which are capable of learning and adapting and whose interactions are non-linear. It is also a system that is constantly evolving and in the process creating new system structures and patterns of behaviour, which makes it difficult to predict and control how the system will respond to changes and interventions. In short, it can be viewed as an ecosystem containing a diverse range of species that is slowly evolving over time and which responds to changes in unexpected ways.
When applying this concept to the informal taxi network of buses and minivans is it possible to identify a number of these elements.
Autonomous, interrelated agents: The primary agents in the system are the taxi operators and the passengers. The taxi operators are autonomous in that they can largely choose which routes and hours they will serve in order to maximise their daily earnings, within the limits of the system (e.g. the available road network). Most taxi operators usually own only one or two vehicles, enabling them to operate independently and thus making them free to identify which routes suit them the most. The passengers are of course free to choose which route, or combination of routes, will get them to their desired destination in the fastest and/or most cost effective manner. They can also choose at which point on the route they will get on and off the taxis. However, they are of course also limited by the availability of taxis. There are also taxi station coordinators and government regulators who play a role in influencing the operation of the system by coordinating the flow of vehicles at major taxi stations and establishing some of the ‘rules of the game’. For this reason the station coordinators and government regulators are much less independent than the taxi operators and passengers but they remain autonomous agents.
System structures and patterns of behaviour (emergence): The informal taxi network is composed of a multitude of operators serving a multitude of routes. The majority of these routes are relatively short return-journeys from one destination to another, and in most cases at least one destination is a major taxi station. Although each of these routes on their own are simple return-journeys, the combination of all these routes together forms a complex transport network that is constantly evolving in response to changes in the system. This makes it difficult to fully understand the scale and coverage of the network let alone gather accurate information on the volume and transfer of vehicles and passengers through the system (i.e. the stocks and flows). An effort was made to produce a map of the main taxi routes a couple of years ago but many of the routes have since changed since then due to major infrastructure projects and construction activities in the city.
Learning and adaptability: The system is adaptive in that the operators and passengers can change their patterns of behaviour in response to changes within the system. This is due to their ability to gather information about changes in the system, learn from other agents, and then respond in the way that meets their individual objectives (i.e. maximising earnings or efficient commuting). For example, taxi operators often modify their routes due to major construction activities disrupting or blocking the regular route. This may initially affect their journey times and profitability until the most suitable alternative route is located but they can still continue to operate relatively unhindered. On the other hand, passengers can change their choice of route(s) in response to such changes if necessary in order to reach their desired destination in the fastest available manner. They can also try to avoid disruptions by exiting the taxi they are traveling on and then walking a relatively short distance to catch another taxi on a different route (I have done this a number of times to avoid traffic jams).
Non-linearity: The question remains as to whether the interactions between the agents in the system produce non-linear outcomes. In other words, whether small changes within the system cause major changes to the overall functioning of the network as a whole (e.g. an increase/decrease in the number of passengers/vehicles, or physical changes to routes). My personal observations are somewhat limited in terms of assessing this aspect. I have certainly observed how construction activities can cause major disruptions and delays to the operation of specific routes. However, the impacts seem to be isolated to one area or route and don’t necessarily cause impacts that spill over onto other routes. In this regard, the system perhaps functions more like a distributed system with a certain degree of redundancy. However, it is likely that there may be tipping points or thresholds which once crossed could cause a major breakdown in the system. For example, delays or disruptions occurring on multiple routes at the same time or an incremental but significant increase in commuter demand.
What insights and considerations can be made from viewing the informal taxi network as a complex-adaptive system?
The observations above indicate that the informal taxi network exhibits many of the characteristics of a complex-adaptive system. However, the question remains as to whether it meets all the criteria, specifically non-linearity, and it would require a detailed study to test this. Nevertheless, some preliminary insights and implications can be drawn from this analysis:
- Despite a lack of planning and coordination the informal taxi network operates relatively effectively and efficiently. Generally the assumption is that well planned and centrally coordinated public transport systems are the most efficient and effective at meeting commuter demand. The case of the informal taxi network in Addis Ababa suggests that some types of public transport can operate effectively and efficiently without such planning and coordination, although the outcomes of the system may be sub-optimal. The main failure of the network (putting aside safety and sustainability issues) is where commuter demand is not being met on certain routes at certain times of the day. The reason for this unmet demand could be the result of a number of factors, such as insufficient information, lack of information exchange between agents, lack of incentives for taxi operations (e.g. profitability of certain routes), or merely the lack of sufficient taxi operators in the system.
- Government efforts to improve the public transport system and planning in Addis Ababa need to take into consideration the role and function of the informal taxi network, in particular its operational dynamics. Currently there are a number of initiatives being implemented to improve the public transport system in Addis Ababa, most notably the construction of a light rail network. At this stage it is not clear if or how the informal taxi network will be integrated with the light rail network (based on the limited available information here and here) although the Transport Policy of Addis Ababa does indicate that intermodal connections will be facilitated by the construction of suitable rail terminals.
The government is also seeking to regulate the informal taxi network to improve safety and ensure routes are being served adequately. This appears to be having positive and negative impacts. On the one hand the regulators have been relatively successful in limiting overcrowding on taxis. One the other hand they have caused delays and confusion at taxi stations by trying to dictate which route a taxi should be operating rather than letting the taxi operator make that choice. A better approach could be to identify where commuter demand it not being met and then finding ways to incentivise taxi operators to serve those route or increase services by government-run buses on these routes. Alternatively they could try to improve the flow of information across the system to make it easier for both taxi operators and passengers to meet their individual objectives. Either way the key point is that developing a better understanding of the dynamics of the system is important before seeking to regulate it or integrate it with other modes of public transport.
I would be interested to hear other opinions on whether the informal taxi network in Addis Ababa does indeed exhibit the characteristics of a complex-adaptive system and also to know if any modelling of informal transport networks has been undertaken either in Ethiopia or elsewhere in Africa. So please share your thoughts in the comments section below.
Adrian Young is an Australian currently living and working in Addis Ababa. He holds a Masters of International Urban and Environmental Management from RMIT University in Melbourne, Australia, and has worked within a number of government agencies and international development organisations in Australia, Ethiopia and South Africa. He has a strong interest in the dynamics and interconnections between human and ecological systems with a particular interest in sustainable food systems and pathways to sustainability. He is the author of Complexia blog where he seek to share his ideas on these issues and how they could be applied in a practical sense.
Read older posts from this section