Travel mode preference in Addis Ababa, Ethiopia: empirical observations

ABSTRACT This research paper analyzes urban travel mode preference in Addis Ababa using a survey distributed to 457 public transit users. The article highlights the details of individual travel and then discusses the implications of the travel patterns for transport policy in the city. The study employed percentages and run multinomial logistic regression to analyze the relationship between mode choice, travel characteristics and socio-economic features of individuals. In terms of mode choice and the influential factors, the analysis shows that income, age, travel expenditure and distance are the determining factors that affect mode choice. It was also found that the privately owned public transport, Taxi (mini-bus) and Higer (midi-bus) form the dominant mode of transport used by closely half percent (41.3%) of the survey respondents. On the basis of our analysis, we suggest that policies have to be devised to make government owned public transport service better in order to support a shift from using low occupancy vehicles to high occupancy vehicles which ultimately solve the problem of congestion and environmental pollution.


Introduction
The role of transport in the spatial and economic development of cities and regions is enormous.Kehinde (2019) in emphasizing the importance of transport stresses the fact that various places of activities within the urban area are geographically separated so a wellmanaged urban transportation is required to overcome the spatial barriers and for cities to function efficiently.
The growing proportion of urban residents coupled with the rate at which urbanization is taking place is creating an unprecedented demand on urban services.As more people shift their living to cities, they require more houses, shops, schools, health centers, roads and public transportation.High travel demand has caused various problems such as traffic congestion, parking difficulties, traffic accidents, public transport inadequacy and environmental problems (Kehinde, 2019;Tsay & Herrmann, 2013).However, most cities have few resources to respond to the magnitude of the change in the urban areas.Many countries attempt to address this issue at a policy level by encouraging greater usage of public transport and nonmotorized transport modes, intensive management of travel demand, traffic management techniques to increase the practical capacity of available road space and improved urban public transport systems (Dhingra, 2011;Pojani & Stead, 2015).Urban transportation and mobility have therefore come to be one of the planning and development issues among urban planners and city managers in developing countries in general (Kehinde, 2019).
Particularly, cities in developing countries have faced difficulties in getting the necessary information to properly plan their urban public transport system.As Turner and Adzigbey (2012) stated the main reason is that most of the developing cities do not have enough budget to collect detailed information necessary for the transportation planning.For example, countries in the sub-Saharan African region have limited data for transport planning, most of the cities do not conduct traffic counting in a regular basis and, except for cities in South Africa, no other city in the sub-Saharan countries has household travel surveys (Ibid).Merkebe et al. (2016) suggested that in order to determine the public transport demand and provide appropriate solutions for a city, an understanding of the travel patterns is vital.
Our aim in this research is to fill the gap in understanding of travel patterns and preferences which is attributed to the limited resources available to conduct household travel surveys.We also examine how the choice of transport options affected by travel and individual features.We then analyze the policy implications of the findings of the paper.These analyses can provide a better basis for policies aiming to enhance transportation access and mobility for urban residents.

Literature review
Urban areas have expanded in relation to overall population growth and have brought challenges on many fronts across the globe.As Motta et al. (2016) indicated urban population is increasing by more than 6% annually in most developing countries.In Africa, the population is over 1.3 billion people, and it is expected to reach 2.5 billion people by 2050 (Arroyo-Arroyo et al., 2021).Consequently, urban mobility has become one of the key challenges for African cities.As a result, leaders and policymakers are expected to provide an enormous amount of infrastructure which is needed to accommodate this growth in urban areas (Tsay & Herrmann, 2013).
In developing countries, public transport is the dominant means of travel for the majority of urban dwellers.In the case of urban transport, Mishra (2019) indicated that public transit is particularly strong and important in developing countries as it plays a crucial role in enabling cities to function as engines of growth.As a result, it is taken as an important solution for narrowing mobility gap and to address issues of social equity.In line with this notion, scholars in China suggested that large cities must implement public transit strategies, make them a priority and establish a public-transit-oriented land use pattern to achieve the goal of urban sustainable development (Cao et al., 2018).It is also believed that the negative impact of car traffic such as noise, pollution, congestion and the risk of accident can be reduced if people are encouraged to use public transport instead of private cars (Holmgrena & Ivehammar, 2020).There is also a consensus that efficient system of public transport can address the urban mobility challenges being faced by many cities of developing world (Mishra, 2019).
However, even if public transportation is relatively inexpensive and efficient, many people still cannot afford it.As Motta et al. (2016) revealed many Indians cannot afford the low fares on public transport.The transit environment in many developing countries is characterized by traffic congestion, delays, mismanagement, mixed traffic, poor coordination among organizations and environmental problems (Mahmud & Rabbani, 2012).Developing countries have several factors in common that contribute to the severity of their transport problems.Overall, population growth and increasing urbanization have led to rapid growth of large cities, which overwhelm travel demand.Public sector finances, in general, are so limited that funding for transport improvements is inadequate (Merkebe et al., 2016;Motta et al., 2016).
There is a substantial literature on urban transport in the developing world, but there are few empirical studies that focus on the transport challenges faced by the urban poor in the developing world (e.g.Mahmud & Rabbani, 2012;Salon & Aligula, 2012;Salon & Gulyani, 2010).Gathering travel information of urban residents is very essential for planning as people with lower incomes are the group that most depends on mass transit to move around in cities (Motta et al., 2016).However, providing reliable and efficient transportation is a key challenge in many developing nation cities (Salon & Aligula, 2012).When cities fail to provide effective public transit or safe spaces to walk or bike, urban dwellers increasingly turn to cars, the most energy-consuming form of transportation (Holmgrena & Ivehammar, 2020;Tsay & Herrmann, 2013).This is also evident in many developing countries that the use of public transport is declining at a fairly constant rate and being replaced by individual transport (Arroyo-Arroyo et al., 2021;Motta et al., 2016).For instance, in Africa, bus operators, private and public owned, have largely been replaced by a fragmented small vehicle paratransit sector (Gwilliam, 2003).
Increasingly over the recent years, particular attention has been paid to urban mobility by governmental bodies, policymakers, transport operators, researchers and user groups.The concern for the sub-Saharan Africa transport policy program is shifting towards improving urban mobility and accessibility (Arroyo-Arroyo et al., 2021).The research community around the globe investigates urban mobility from various angles in an attempt to find new innovative tools, emerging ideas and approaches (Tyrinopoulos & Antoniou, 2013).This paper is also among the endeavors which aim to gain insight on urban travel features.Therefore, the insights gained in this paper could be used to determine policy measures to reduce transit challenges being experienced/faced by users.

Transit profile of Addis Ababa
Addis Ababa is the political and commercial center of Ethiopia with a population of 5 million (UN, 2018) and with a dominant contribution, 50%, to National Gross Domestic product of the country (Feyissa et al., 2018).
In Addis Ababa, government and the private sector are service providers in public transportation.The modalities of public transport mobility in Addis Ababa involve road and rail-based transit.The road-based transit mainly comprises Anbessa buses, Sheger buses, Higer (midi-buses), Taxi (Minibus) and Salon taxis (see Figure 1).The two-government operated public transport enterprises, Sheger and Anbessa, provide standard bus service in the city (AACRA, 2014).
Based on data from AACRA (2020), Anbessa has 124 routes add up to 1,528 km route network, most of the routes are overlapping at the center of the city.Sheger operates on 48 routes with 547 km route network.The LRT operates two routes with a total length of 34 km; Line 1 (East -West) operates from Ayat to Torhailoch with 17 km and 17 stations and Line 2 (North -South) operates from Kality to Menelik II Square with 16.9 km and 22 stations.In the case of Mini-bus (Taxi), the numbers of routes are 1, 529 with 1,358 km network coverage (see Table 1).Higer also operate on 37 routes (Davis, 2022).However, the large proportion of the routes is found in the central areas.For instance, close to 70% of Taxi operating routes found in the inner sub-cities such as Addis Ketema, Kirkos, Arada and Lideta (Davis, 2022).However, the service is overcrowded, operated in mixed traffic and used old, unreliable and inefficient vehicles (Mulu, 2015).People use walking as a mode of transport in response to transit challenges such as absence of service and inability to afford the existing means of transport (Gebremeskel et al., 2022).
Despite the existence of different public transportation services, the city experiences a huge problem in providing efficient transportation services to the city's residents.The city's growth in economy, geographical area and population, brings urgent attention and needs additional mass transport service provision.Hence, improving urban public transport is among the issues that have been given attention to overcome mobility challenges.It has been stated that the future development direction of the city shall consider the transport infrastructure and mass transport service provision (MOT, 2011).In realization of this, a great effort is being made by the government to enhance mass transport by introducing Light Rail Transit (LRT) and Bus rapid Transit (BRT) services.The operation of Light Rail Transit service started in 2015.The city's first bus rapid transit is also under construction and in its final stages.

Data and method
A quantitative research approach was adopted for this study.The data were gathered through a survey questionnaire distributed to public transit users on main transit stations.The selected stations include Torihalioch, Stadium, Megengna, Ayat, Merkato, Kality, Menelik square, Piazza and Mexico square.In addition, transit stations that are found in peripheral areas (mainly condominium areas) were purposively selected as study sites.These include Jemmo, Yeka Abado, Tafo, Tullu dimitu (see Figure 2).A convenience sampling technique was used to reach 457 respondents.

URBAN, PLANNING AND TRANSPORT RESEARCH
The socio-economic attributes and travel characteristics gathered through a survey questionnaire include travel expenditure, distance, income, age and gender.The collected data were analyzed using percentage, and we also run multinomial logistic regression to estimate the influence of socio-economic attributes of individuals and travel characteristics on travel mode choice of respondents.
In Addis Ababa, people may use either privately owned or government owned or a combination of them for their daily mobility.Respondents are asked the specific travel mode they use and here for the purpose of multinomial logistic regression analysis all of the government owned modes are aggregated into a single category, public transport (including Anbessa bus, Sheger bus, Light rail transit).Likewise, private owned transport modes include Higer (midi-bus) and Taxi (mini-bus).The paper considered the most frequently used means of travel.Since the focus of the research is on public transit system that is frequently used, walking, bicycle and other forms of transport are not included in the analysis.

Result
According to the survey results, the share of female and male respondents is 52.3% and 47.7%, respectively.In terms of employment status, close to 60% of respondents work for government and privately owned companies that have daily fixed working times and then required to commute every day.More than 25% of survey respondents are engaged in personal business, both formal and informal.Few of them, less than 5%, are housewives and pensioner who did not have to travel every day to work, and the remaining are students.
The data in Table 2 also show that more than half of the respondents are in the 26-45 age groups, suggesting that they are working age groups who are required to commute on a daily basis.In terms of income, more than half of the respondents have a monthly income of 4,000 to 10,000 Birr, while 23% earn less than 4,000 Birr each month.The remaining 20% earn more than 10,000 Birr each month (see Table 2).
With regard to educational level, those who cannot read and write represent a very small percentage, 0.4%.Those who have a lower educational attainment (who can read and write) also account 2.8%.This actually signals the increasing efforts being undertaken by the government in increasing literacy rate across the country.Those who attend primary school education (1-6) represent 2%, junior school (7-8) 1.8% and high school (9-10) 11.6%.Preparatory school (11-12) takes 7.2%.TVET (Technical and Vocational Education and Training) certificate accounted 10.7% while diploma holders accounted 21.7%.Bachelor's degree holders represent 32.4% and 9.4% of the respondents are Master's holders and above (see Table 2).

Travel features
Transport is an indispensable component of human activities worldwide.The movement of people, goods and services within the urban area has been a vital aspect of the city life.Various places of activities within the urban area are geographically separated and so transportation is required to overcome the spatial barriers and for cities to function efficiently (Kehinde, 2019).
People use a combination of modes of transport for their daily travel.Among the survey respondents, 36% reported that Taxi (mini-bus) was their most frequent mode of transport.Government owned Anbessa and Sheger bus are also the most common transport mode in Addis Ababa, used by nearly 30% and 25%, respectively.The least transportation mode used by respondents was Higer (midi-bus) and Light Rail Transit, which share 5% and 4%, respectively (see Figure 3).As can be seen from Table 3, there is no significant difference among male and female in their choice of forms of transport.Short distance travelers (below 5 km) prefer to use privately owned modes (e.g.Taxi) while long distance travelers (6 and above) prefer to use public owned modes and a combination of public and private owned modes.Likewise, the majority of respondents (60%) in the age group above 55 prefer to use government owned modes (Sheger, Anbessa, LRT) and young age groups (below 35) prefer to use a private owned mode and a combination of private and public over a public owned only.In terms of income, 45% of respondents in a lower income category (below 4000) rely on government owned modes for their daily mobility.Only 12% of them use private owned modes.Respondents in the higher income category (Above 8001) highly rely on a combination of private and public owned modes.Compared to the government owned modes which share only 12%, 30% of respondents in higher income category use private owned modes.
Figure 4 represents the share of trip by purpose on weekday and weekend.As it is depicted, work trips take the largest share on weekday, which is 45%, followed by school trip and trip for taking children to/from school, 19% and 17%, respectively.Shopping and social trips (recreation, family visit, worship) represent a share of less than 10%.Unlike  weekday, social trips (recreation, family visit, worship) represent 60% while shopping trips account 25% on weekend, whereas work, pension and school trips that occur on weekend represent not more than 5%.
The data in Table 4 shows that Taxi mode is found to be the desired choice for all trip purposes.To mention some, more than 50% of respondents indicated that Taxi mode is their preferred mode for recreational and medical trips while more than 30% preferred to use Taxi for school, to take children from/to school, work, shopping and personal business.
A significant number of respondents preferred to use Anbessa and Sheger modes for their daily travel.For instance, close to 40% of respondents prefer to use Anbessa bus for visiting family, friends and attending social events while 23% preferred to use Sheger bus.For work trip, Anbessa is a preferred mode for 24% of respondents and Sheger for 13.6%.Similarly, Sheger is also found to be the preferred mode for recreational trip by 14% of respondents while 3% goes to Anbessa bus.The table also depicts that the share of Sheger and Anbessa buses accounts less than 5% for school, to take children from/to school and medical trips.Shopping trips to be made by Anbessa and Sheger buses account 13% and 10%, respectively.In the same fashion, 10% of respondents preferred to use Anbessa and Sheger buses for religious trips.
Among all users, the least preferred mode of transport for all trip purposes is Higer followed by Light Rail Transit.Higer accounts less than 5% for all trip purposes except work which represent 5% while Light Rail Transit represents not more than 7%.
Table 5 highlights the factors that influence the choice of a particular mode of transport.Of these factors, speed, ease of availability and frequency of service are the most important factors to Taxi users, which represent close to 20%, whereas distance to station, operating hour, convenience represent in between 10% and 15%.The least important factors in choosing Taxi mode are cost and comfort, 3% and 7.4%, respectively.For Higer mode choice, cost, ease of availability and distance to station are found to be important factors as they represent 30%, 36% and 21%, respectively.Respondents do not consider convenience, comfort and speed in choosing this mode while 3% of them consider frequency of service.
The data also show that cost is the most determining factor for Anbessa (33%) and Sheger (28%) bus selection.Ease of availability followed cost with a value of more than 20% for both buses, Anbessa and Sheger.Similarly, distance to station and operating hour are also influential factors in choosing Anbessa and Sheger buses representing more than 10%.The least considered factors for Anbessa bus are speed (0.4%), comfort (0.1%) and convenience (0.8%), while frequency of service accounts 7%.Likewise, for Sheger, bus speed, comfort, convenience and frequency of service represent a small share, 6.9%, 2.4%, 7% and 9%, respectively.
Looking at the data in Table 5, in choosing Light Rail Transit speed, cost and frequency of service are the most important factors representing 28%, 26% and 20%, respectively.Among other factors, convenience and operating hour account 8% and 7% and the remaining ease of availability and distance to station are found to be the least contributing factors with less than 5%.

Results of multinomial logistics regression
The regression analysis aimed to estimate the influence of individual and trip attributes on the public transit mode choice.The likelihood ratio test proves that the independent variables such as age, income, travel expenditure, and distance significantly contribute to the model (see Table 6).
From the results of multinomial regression analysis, income, age, travel expenditure and distance travelled have a significant impact on the mode choice of respondents.As age, income, distance and travel expenditure increased by 1 unit, the probability of choosing public owned mode relative to privately owned mode is more by 1.057, 1.000, 1.100, 1.081 times, respectively.However, gender of the respondents was not statistically significant.In the combination of public and private owned category, income, distance and travel expenditure, but not age and gender, had a significant impact on the choice of a combined mode relative to a privately owned mode.For instance, as income, distance and travel expenditure increased by 1 unit, the odds ratio/probability to choose combined public and private modes relative to only a private mode was more by 1.000, 1.097, 1.180 times, respectively.

Desired improvements on public transport service
Transport service provision in most developing countries is unable to cope with increasing demand.Thus, people have to face great inconveniences, delay, cost, frustration, accident and loss of life.Respondents were asked their opinion about priority on public transit improvement requirements in the city.The question produced an important message on public transport service improvement requirements.
As it is summarized on Figure 5, the first suggested area for improvement is availability (35.4%) followed by safety and security (22.5%), affordability (21.4%) and comfort (20.6%).Obviously, public transit users aware of the need for the improvement of the above-mentioned service parameters.In the case of Addis Ababa, the problem related to public transport service is the inaccessibility of modes.It is very common to see longer queues on public transit stations.

Discussion
The result of the paper revealed that city dwellers use a combination of public transportation modes for their daily travel.Salon and Gulyani (2010) also reported that the vast majority of Nairobi residents use public transport as their most frequent means of travel.Similarly, in Dakar, the mobility of people is ensured in part by the large-bus services and by the paratransit service (Merkebe et al., 2016).
Although people use a combination of modes, Taxi mode is found to be the desired choice for all trip purposes.For instance, compared to the government owned public transit modes, respondents in higher income category use private owned modes.Salon and Aligula (2012) also revealed that in Kenya those who can afford to pay for transportation choose public transport service provided by Matatus, whereas respondents in a lower income category rely on government owned modes.Hernández (2017) also revealed that low-income groups in developing world depend on public transportation.The result of our study also showed that Taxi transport is mainly used by short distance travelers, while long distance travelers choose to use government owned modes, Anbessa, Sheger and Light Rail Transit.This is attributed to the fact that government owned modes have long distance coverage and cost advantage.
The most striking finding from multinomial regression analysis is that with increasing income people prefer to use public owned modes and a combination of public and private over privately owned modes.This may be attributable to the fact that the impact of congestion has become common for all mode users in the city.Thus, people may tend to use any mode of transport at their convenience.As it is indicated by Mishra (2019) the majority of population in developing world cities belong to the poor, low income and middle-income groups.In the case of Addis Ababa, low-income and middle-income groups are working groups and tend to live in the outskirts of the city where affordable housing is available.For this reason, they may prefer to use public buses for long distance coverage and lower cost advantage.The other reason is that the city transport officials have been striving to improve the service provision of public buses by providing exclusive lane which has improved their speed.In addition, the rising cost of living coupled with increasing price of petrol which has influenced cost of transport has resulted in a shift from using high cost private owned modes to affordable government owned modes.
In choosing a travel mode for a particular trip, people consider a number of factors including cost, time, comfort, safety and availability (Mahmud & Rabbani, 2012;Salon & Aligula, 2012).Of these factors, speed, ease of availability and frequency of service are the most important factors to Taxi users.The least important factors in choosing Taxi mode are cost.Like Taxi transport in Addis Ababa, the Matatu paratransit system in Kenya is unaffordable for the majority of the poor (Salon & Gulyani, 2010).
The least considered factors for government owned buses (Anbessa and Sheger) are speed, comfort, convenience and frequency of service.In Addis Ababa people may not prefer to use government owned modes as they are less convenient and comfortable.As it is indicated by Tyrinopoulos and Antoniou (2013) crowding is the factor that most discourages people from the use of public transport.The data also show that cost is the most influential factor for government owned modes (Anbessa, Sheger and Light Rail Transit) selection.This finding is consistent with the case in Dhaka where fare per trip is reported to be an important factor for choosing public means of transport (Mahmud & Rabbani, 2012).
Respondents suggest desired improvements on public transit service provision.The areas suggested for improvement include availability, followed by safety and security, affordability and comfort.In most developing countries, the burden of transport expenditure on poor people is very high, excluding them from work and social services.A study also found that many Indians cannot afford the low fares on public transport (Motta et al., 2016).Similarly, in Kenya, it was found that the transit system is physically accessible for poor residents but affordability remains a large constraint (Salon & Gulyani, 2010).In terms of safety and security, women face violence and sexual harassment while using crowded buses (Ceccato, 2017;Gwilliam, 2003).

Conclusion and recommendation
Africa's population continues to grow at a rapid rate.As a result, urban mobility has become one of the key challenges for African cities. Very large number of people need to travel and its very common to see longer queues.The problem is also exacerbated by the rapid increase in motorization and the rise of informal transport based on large number of small vehicles.Addis Ababa is not an exception of this reality.
This research paper discussed the features of urban travel in Addis Ababa, Ethiopia.The study used data collected through questionnaire from 457 public transit users.Convenience sampling was used to sample respondents on major transit stations, while purposive sampling technique was employed to select major transit stations.The selected stations include Torihalioch, Stadium, Megengna, Ayat, Merkato, Kality, Menelik square, Piazza and Mexico square.In addition, transit stations that are found in peripheral areas (mainly condominium areas) were also included as study sites.
In terms of mode use, a significant number of respondents use Taxi and government owned buses, Anbessa and Sheger.The result of multinomial regression analysis showed that with increasing income people prefer to use public owned modes and a combination of public and private over privately owned modes.The most obvious message from this research is that private owned public transit, Taxi, continues to dominate urban travel among every segment of the study population in Addis Ababa.However, the combined share of Anbessa and Sheger bus usage (55%) by far exceeds the share of Taxi usage (36%) (see Figure 2).
The results drawn from the analysis can be used for transport planning and policies.For instance, despite the presence of various service limitations in public transport, respondents still use public transport over privately owned modes.Such findings are quite useful for policymakers to apply appropriate measures to enhance the service provision of government owned public modes and improve its attractiveness.Therefore, the overall policy implications of this analysis are that the small occupancy vehicles (Taxi) cannot be taken as a main strategy for solving or meeting the mobility requirements of urban dwellers.Government owned public transit (high occupancy vehicles such as buses) should be taken as area of improvement.At last, the knowledge presented in this paper could be used to determine measures to reduce long-term dependency on paratransit system in favor of more sustainable modes of transport such as buses and rail-based transit.The literature also suggested to enhance the affordability and desirability of public transport (Salon & Gulyani, 2010) as the privately owned para-transit system has serious operational and other broad level limitations such as the creation of congestion and environmental implication.

Disclosure statement
No potential conflict of interest was reported by the authors.

Figure 5 .
Figure 5. Desired improvement requirements on public transport service.

Table 1 .
Public transit routes and coverage.

Table 2 .
Socio-demographic profile of respondents.

Table 3 .
Cross tabulation of mode choice by personal and trip attributes.

Table 4 .
Mode preference and trip purposes.

Table 5 .
Reasons for choosing a particular mode of transport.

Table 6 .
Results of multinomial logistic regression.