Hybrid urban transport systems in developing countries: Portents and prospects

Research in Transportation Economics 39 (2013) 121e132

Contents lists available at SciVerse ScienceDirect

Research in Transportation Economics journal homepage: www.elsevier.com/locate/retrec

Hybrid urban transport systems in developing countries: Portents and prospects Pablo Salazar Ferro*, Roger Behrens, Peter Wilkinson 1 Centre for Transport Studies, Department of Civil Engineering, University of Cape Town, Private Bag X3, Rondebosch 7701, South Africa

a r t i c l e i n f o

a b s t r a c t

Article history: Available online 30 June 2012

Some South African cities have initiated public transport transformation projects which, in most cases, ultimately envisage the in toto replacement of paratransit operations with formalised BRT systems. There are two likely outcomes: (1) complex negotiations with existing operators and budget constraints will result in the in toto transformation occurring over an extended period of time; or (2) in toto transformation will simply not occur. In either case, cities will depend, for decades, on a ‘hybrid’ public transport system that combines both formal and paratransit operators. This paper presents a case for policy recognition of hybrid systems, and explores how such systems might best be managed. The following categories of hybrid public transport systems are explored through case studies: (1) transformative processes in which paratransit operators are to form or assimilate into companies to operate new services, but this incorporation has proved difficult to complete and the operational and regulatory frameworks remain unchanged; and (2) transformative processes that, from the outset, anticipated a hybrid system and designed the outcome accordingly. A third category of hybrid transport systems, defined as transformative processes that have been amended following a realisation that in toto transformation is unattainable, is also introduced and discussed. The paper concludes by tentatively drawing lessons for appropriate public transport regulation, particularly with respect to Cape Town’s transformation project. It is argued that a review of the current national regulatory framework is required to enable possible project modifications that acknowledge system hybridity. It is suggested that regulatory frameworks that accommodate the likely hybrid nature of public transport system outcomes have greater prospects of success than frameworks that do not. Furthermore, it is argued that contextually appropriate and successful public transport transformation projects do not necessarily require the in toto substitution of incumbent paratransit operators, and that they can be integrated with, and complement, formal services. Ó 2012 Elsevier Ltd. All rights reserved.

Keywords: Paratransit Hybrid public transport system Bogota Accra Cape Town

1. Introduction Public transport systems in South African cities reflect the historical spatial planning policies that shaped urban territories. Apartheid settlement policies dictated that large pockets of lowincome inhabitants would be located at the peripheries of cities. Initially, rail- and road-based services transported workers in and out of employment zones. However, starting in the 1970s, the minibusetaxi sector began supplanting all existing modes and became the dominant form of public transport in most cities (Clark & Crous, 2002). During the 1980s, the minibusetaxi industry experienced rapid growth (Browning, 2001) and its share of modal

* Corresponding author. Tel.: þ27(0)21 650 4756. E-mail address: [email protected] (P. Salazar Ferro). 1 In memoriam: Peter Wilkinson passed away on 10 August 2011, before the finalisation of this paper. 0739-8859/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.retrec.2012.06.004

splits increased accordingly. From1998,2 the two principal modes of public transport e passenger rail services and scheduled bus services e showed signs of significant service quality decline (such as overcrowding, safety, unreliability, amongst others), while ‘paratransit’3 modes were unable to offer a higher quality of service (Wilkinson, 2008). Contemporary policy discourse holds that the increasing dominance of minibusetaxis has gradually undermined other existing modes and has been partially responsible for the decline of the public transport system.

2 In September 1998, the initial proposals of a first ‘post-Apartheid’ national transport strategy were published in the Moving South Africa document. 3 ‘Paratransit’ is defined in this paper as unscheduled and unplanned urban public transport services generally operating outside the formal regulatory framework. These services often start as illegal services and acquire legality as they become more established in the city. Services are road-based, with vehicles ranging from multi-seat motor-tricycles to full size conventional buses. Vehicles used often lack minimum safety standards and are not necessarily periodically maintained.

122

P. Salazar Ferro et al. / Research in Transportation Economics 39 (2013) 121e132

Symptoms of dysfunctionality in South African public transport systems include: the mode share decline of (rail- and road-based) ‘formal’4 services; violent competition in the market on overtraded routes; the poor quality of service provided by ‘paratransit’ modes (most notably aggressive driver behaviour and the lack of roadworthiness of vehicles); and weak law enforcement and regulation (Schalekamp & Behrens, 2010). The South African National Department of Transport (NDoT), in response to these problems, initiated different projects that were aimed at modifying the existing public transport system. The Taxi Recapitalisation Programme of the early 2000s and the various Integrated Rapid Public Transport Network (IRPTN) projects of the late 2000s were the cornerstones of the proposed transformation of public transport systems. Partly in response to the imperative of preparing the transport systems of host cities to better accommodate the 2010 FIFA World Cup events, substantial sums of money were made available in 2007 through the Public Transport Systems and Infrastructure Grant (PTSIG) and the NDoT to facilitate the planning, design and implementation of IRPTNs in different cities.5 Generally, the proposed projects incorporated a network of bus rapid transit (BRT)-type corridors modelled along the lines of the systems introduced in Latin American cities such as Bogota and Curitiba, with the former being the most recent and publicised example (Lleras & Pereira, 2009). In most cases, the IRPTN proposals have involved the wholesale or phased total replacement and transformation of the existing mixed ‘formal’/’paratransit’ system by a comprehensively regulated and formally operated system. This paper explores the likely outcomes of current policy objectives of in toto paratransit replacement and wholesale transformation. The paper is divided into five sections. The following section describes national public transport policy in greater detail, and illustrates the manifestation of this policy in the case of the city of Cape Town. Section 3 reviews the outcomes of the pursuit of similar policy objectives elsewhere. Section 4 draws lessons from these experiences, and Section 5 concludes with a discussion of future research needs.

conditions for interim phases between the initial planning and full implementation of all phases of IRPTN systems in a particular city: Even though the early Phases of the IRPTN will not cover all existing bus and minibus routes, the long-term IRTPN operation system planning commences in year one and the existing routes to be eliminated or modified should ideally not be planned according to the NLTTA RatPlan or OLS process. The IRPTN should replace these, even for areas not covered in the initial Phases. This is because neither the RatPlan nor the OLS were Operation Plans, nor could they effectively rationalize or optimise the services as has been demonstrated by the IRPTN Operational Plan process. Thus, it is suggested that in the interim, the Interim Contracts should remain in place on a monthly renewal basis and the moratorium on new licenses should remain, affording the greatest flexibility for the phased implementation of the full IRPTN Operations Plan. Source: National Department of Transport, 2009, page 42. The proposed interim condition avoids recognition of a system composed of ‘paratransit’ services and ‘formal’ services. Hence, the management and regulation of the interface between these two types of services are not addressed as no regulatory mechanisms are created for such conditions. The implementation of IRPTN systems was expected to be quick and wholesale, but in light of new information concerning phases for all six metropolitan areas involved in CITPs, the interim period will be significantly longer than initially contemplated.6 Cities will then be forced to follow and apply regulatory frameworks not designed for transitional periods during an interim period lasting for decades in most of the cases. Cities could thus experience difficulties in responding to new patterns of passenger demand associated with changing urban conditions relying solely upon monthly interim contracts and unable to issue new operating licenses. The public transport policy dichotomy was created with the introduction of IRPTN Operational Plans. Two different frameworks have to be implemented by the national government and the provincial governments, while local governments should apply one of the two existing frameworks. Cape Town, as one of the largest urban municipalities, is included in the cities that are to apply an IRTPN.

2. Cape Town’s public transport system 2.2. A mixed but dysfunctional system 2.1. South Africa’s policy dichotomy The South African public transport policy environment presents a dichotomy in so far as two sets of policy prescriptions are applied to two types of cities: those with IRPTN or BRT projects; and those without. Depending on the type of city, local governments need to apply and uphold different frameworks. Cities not concerned with IRPTN plans are required to continue to apply earlier regulatory frameworks such as Rationalisation Plans (Ratplans), Operating Licensing Strategies (OLSs) and provincial bus contracting. On the other hand, cities involved in IRPTN planning, need not apply earlier regulatory frameworks. This group of cities should replace previous frameworks with an IRTPN operational plan. The NDoT in the planning guidelines for Comprehensive Integrated Transport Plans (CITPs) explains the

4 ‘Planned’ or ‘formal’ services are defined in this paper as public transport that operates within a formal regulatory framework providing scheduled, planned or frequency controlled services on fixed routes. Vehicles in these networks are typically periodically maintained and they respect minimum safety standards. The network often encompasses rail-based systems and road-based systems that normally use high capacity vehicles. 5 The key references to the transformative process are the Public Transport Strategy (NDoT) and the Public Transport Action Plan (NDoT) of 2007.

Cape Town’s public transport system has some of the more important features of South African cities but, at the same time, it is an unusual system in terms of its modal split. There is an observable duality in the general transport system as lower income households located in distant zones from the employment areas are the main users of public transport, while middle- and high-income households in more convenient locations generally depend upon private cars to fulfil their mobility needs (Wilkinson, 2008). In terms of public transport services, the key feature of the system is the role of the rail-based services that can be described as the backbone of the transport system in Cape Town on the grounds of passenger numbers. Table 1 shows the size of the public transport passenger market in 2005 for the Cape Town metropolitan area. In practical terms, public transport ridership numbers depend on the affordability and the accessibility of the three main modes.

6 The Public Transport Action Plan of 2007 requires 12 cities to implement catalytic IRPTN projects. These 12 affected areas are: Johannesburg, Tshwane, Cape Town, Nelson Mandela Bay, eThekwini, Ekurhuleni, Buffalo City, Mbombela, Mangaung, Polokwane, Rustenburg and Msunduzi. The strategic phasing presented in this document defines three phases, two of which concern the 12 cities. Phase I (2007e2010) involved the implementation of catalytic IRPTN projects. Phase II (2010e2014) required the completion of entire IRPTN networks.

P. Salazar Ferro et al. / Research in Transportation Economics 39 (2013) 121e132 Table 1 Cape Town metropolitan daily public transport passenger trips in 2005, by mode.

Table 2 Evolution of Cape Town’s public transport passenger market share.

Passenger rail Scheduled bus Minibusetaxi Total services services (GABS) services passenger (Metrorail) trips Daily quantity of 601,900 passenger trips (Percentage share) (54%)

197,400

332,400

1,131,700

(17%)

(29%)

(100%)

Source: City of Cape Town, 2005.

Subsidised rail and bus services typically have lower fares than minibusetaxi services, but the latter provides more frequent and conveniently located services (Wilkinson, 2008). Furthermore, by nature, rail services provide less flexible services and are better suited for line haul travel within an urban region. Subsidised road services, in the form of ‘formal’ buses, are more flexible but provide infrequent services (Golden Arrow Bus Services (GABS) averages one bus every 15 min in peak hour, while train frequencies range from 3 to 8 min during the same period (City of Cape Town, 2005)). Minibusetaxi services, on the other hand, are extremely flexible and demand responsive. These ‘paratransit’ services quickly respond to user needs and can easily adapt to dynamic patterns of demand (Clark & Crous, 2002). Problematically, in some routes, minibusetaxi services and ‘formal’ bus services compete for passengers (McLachlan, 2010). In that sense, minibusetaxi operations have taken on longer routes that, some argue, are not suitable for small size vehicles (Browning, 2001). Additionally, the fleet’s small vehicles have indisputable advantages when travelling through the narrow roads of informal settlements and thus they become the main form of transportation in these areas (Clark & Crous, 2002). These modal advantages and disadvantages, as shown in Table 2, have resulted in an increase of the minibusetaxi’s share of the public transport passenger market and a resulting decline in the share of ‘formal’ services. In terms of the ‘formal’ and ’paratransit’ shares, Table 2 shows that overall ‘paratransit’ services have been increasing their importance when compared to the total share of ‘formal’ modes (rail- and road-based). They are now responsible for almost onethird of public transport trips in Cape Town, as opposed to onetenth of trips in 1987. This percentage is considerably smaller than the ‘paratransit’ share of other South African cities,7 but it is a growing element of Cape Town’s public transport system. A key problem with the situation in Cape Town, as is the case of other South African and developing world cities, is the lack of integration between modes (Wilkinson, 2008). Fare structures are different for every mode and under prevailing operating arrangements there is no possibility for a ticket combining rail and road services in a single trip; or for a ticket combining the two types of road-based services. Regulation also remains a challenge. A complicated mix of government spheres8 and levels of involvement are a major obstacle. While the National Government, through its Department of Transport, regulates the parastatal rail company’s services, the Provincial Government is responsible for regulating ‘formal’ private bus services. Minibusetaxi services, on the other hand, are partially regulated by the Provincial Operating License

7 According to the National Household Travel Survey Technical Report of 2003, for all work journeys in metropolitan areas, trains account for 11.3% of all trips, scheduled buses account for 8.7% and minibus taxis for 29.1% (National Department of Transport, 2005). This is translated into public transport market shares of 23.0% for train services, 17.7% for scheduled bus services and 59.3% for minibusetaxi operations. 8 The three spheres of government associated with transport planning in Cape Town are the National sphere, the Provincial sphere and the Local sphere (which can be understood as a metropolitan sphere).

123

Year

Passenger rail services Scheduled bus services (GABS) Minibusetaxi services

1987

1998

2000

2005

65% 24%

62% 18%

62% 14%

54% 17%

11%

20%

24%

29%

Source: Grey, 2006 and City of Cape Town, 2005.

Board in association with the Metropolitan Government, while also having self-regulatory bodies in the form of associations and ‘mother bodies’.9 In sum, three spheres of government are involved in public transport regulation and integration remains difficult to achieve under these circumstances. The integration problem is not only reflected in terms of fares and regulation, but also in terms of coverage. In any one area of the city, public transport systems are mixed but operate independently. Along various corridors ‘formal’ buses and ‘paratransit’ services compete in the market, with detrimental effects. The scheduled bus network has been characterised as a dispersed one (Clark & Crous, 2002) and the competition with minibusetaxi services exacerbates this situation by making some routes unviable. Rail services, on the other hand, have only limited coverage (Grey, 2006) and are not flexible enough to respond to new demands; consequently, their importance within the public transport system has been eroded. Furthermore, there has been a lack of investment (McLachlan, 2010) that hinders the network’s ability to respond to transport demands even along established rail lines. 2.3. Proposals for systemic transformation The problematic lack of integration and the growing importance of unregulated ‘paratransit’ were two of the main issues motivating proposals for systemic transformation. Different documents10 published between 2006 and 2007 depict a model of an integrated network of public transport services. For many South African cities, and Cape Town in particular, the integrated public transport system envisioned by these documents included rail routes, and new BRT corridors including trunk and feeder services. In this sense, BRT systems were to be the catalyst for major overhaul of the public transport system. Within the envisioned model, the BRT trunk corridors would complement existing rail lines as backbones of the system. ‘Paratransit’ and ‘formal’ operators would be integrated largely as new companies contracted to operate trunk and feeder services. In that model, feeders would be supplemented by non-motorised modes. The proposed system is largely based upon the experience from Bogota, Colombia (Schalekamp & Behrens, 2010). The BRT corridors to be implemented broadly follow a radial pattern focused on the city centre, with axes radiating to the north and southeeast (Phases I and II, respectively) and axes radiating to the east (Phases III and IV, respectively) (as shown in Fig. 1). Most corridors lead to the city centre, thus supporting an important residence e employment destination and also creating a significant

9 ‘Mother bodies’ are officially unrecognised associations of minibusetaxi operators. They usually cover a particular geographical region and encompass local and long distance operations (Grey, 2006). 10 The Local sphere of government produced the Public Transport Plan of 2006, the Integrated Transport Plan for the City of Cape Town 2006e2011 of 2006 and the Public Transport Implementation Framework of 2007. The National sphere produced the National Land Strategic Framework of 2006, the Public Transport Strategy of 2007 and the Public Transport Action Plan Phase I: 2007e2010 of 2007.

124

P. Salazar Ferro et al. / Research in Transportation Economics 39 (2013) 121e132

node connecting different corridors. This configuration essentially reproduces the rail network. Currently, suburban rail lines radiate from the city centre and, combined with major arterial roads, define main activity corridors (Wilkinson, 2000). As presented in the 2007 Public Transport Implementation Framework, Cape Town’s proposal calls for the implementation of a full-specification BRT11 system (City of Cape Town, 2007). Therefore, the proposed trunk and feeder system requires infrastructural and fare collecting components. The project, for its initial phase, envisions dedicated busways in the median, elevated stations (to allow for level boarding of high-platform buses) and reinforced pavements. In terms of the fare collection system, it is expected that pre-boarding ticket control and smart card technologies will facilitate trunk-feeder integration in the final system. The BRT proposal includes four phases as shown in Fig. 1. The Public Transport Implementation Framework document of 2007 proposed the implementation of the first phase over a period of two to three years. This proved to be unachievable. More recent documents indicate a timeframe of 10e12 years for the entire network (McLachlan, 2010), but this has also been questioned. The most recent proposed timeframe of 20 years indicates the following dates for the different phases (Fortune, 2011): -

Phase Phase Phase Phase

I, complete implementation in 2013; II, 2018e2020; III, 2020e2022; IV, 2025e2030.

Currently, the transformation process is focused on the first phase. Phase I concerns a corridor lacking passenger rail services while also having a relatively low number of minibusetaxi operator associations when compared to other corridors. The lower quantity of associations offers a less complicated engagement on paratransit ‘formalisation’ (Schalekamp & Behrens, 2011). This initial phase is confined to a single BRT corridor with multiple feeder services, connecting the employment area of the CBD to the northern part of the city, a largely residential area. The City of Cape Town has defined various intermediate phases (initially Phase 1A and Phase 1B) to fully implement the public transport corridor along this axis. In May 2011, after various delays, initial services for Phase 1A were launched. To be able to start the services, operators were divided into three different companies that have a particular role within the geography of the network. Two of those companies are composed of former ‘paratransit’ operators, with the third company being GABS. Currently all operators are working under an interim contract scheme, to be renegotiated after 10e12 months and converted into negotiated 12-year operating contracts (City of Cape Town, 2011). 2.4. Obstacles in the transformational process As with other South African cities, one of the main difficulties in public transport transformation initiatives has been the integration of ‘paratransit’ operators. The proposed complete substitution of ‘paratransit’ services in Cape Town and other South African cities is a controversial issue. The transformational process offers two distinct possibilities for ‘paratransit operators: their inclusion in the ‘formal’ system as new ‘formal’ operating companies grouping various operators; or a withdrawal from the system with

11 Full-specification BRT systems are defined in this paper as road-based systems consisting of co-ordinated trunk and feeder services, high capacity trunk vehicles (articulated or bi-articulated buses), protected and reinforced exclusive bus lanes, clearly designated stations with level boarding, pre-boarding automated fare collection systems, and real-time conditions information.

Fig. 1. Proposed phases for BRT implementation in Cape Town. Source: City of Cape Town, 2010.

compensation. However, many ‘paratransit’ operators have expressed concerns over their integration into a ‘formal’ system (or their eventual withdrawal from the general public transport system). A key issue is the organisation of the ‘paratransit’ industry. Its fragmented ownership represents an important hurdle in the transformational process. Current ownership structures and business models demand a prior formalisation in order to be included in the newly planned systems (Schalekamp & Behrens, 2009). A reorganisation of the entire industry is thus required. However, the highly fragmented structure coupled with multi-layered associations makes the task difficult. Furthermore, associations are loosely formed (Browning, 2001) and, as is the case with ‘mother bodies’, they have difficulties claiming they are representative of the sector. City governments have therefore encountered understandable challenges in engaging ‘paratransit’ operators in their proposed projects. According to city officials, in Cape Town, negotiations with minibusetaxi operators started in 2008. However, the ‘paratransit’ sector claims there was a lack of consultation. The policy ambition of a relatively quick and wholesale transition from ‘paratransit’ operators to ‘formal’ operators has thus been described as problematic (Schalekamp & Behrens, 2010). Many current operators are sceptical about their prospects in the planned system, and fear the unknown. The minibusetaxi industry would be forced to relinquish control over their operations and form new, less autonomous, entities. By giving up their operating licenses, minibusetaxi operators are forced into an all-or-nothing situation, where failure of the project would result in loss of livelihood with no fall-back option. It is understandable that ‘paratransit’ operators have concerns. A direct relationship between assimilation into the

P. Salazar Ferro et al. / Research in Transportation Economics 39 (2013) 121e132

new ‘formal’ system and a subsequent universally improved and sustained livelihood is uncertain. During protests in 2010, ‘paratransit’ operators stated that the implementation of new BRT-based systems and the integration of ‘paratransit’ could translate into a loss of employment opportunities within the minibusetaxi industry (Schalekamp & Behrens, 2010). Furthermore, even if official proposals claimed that there would not be job losses e a difficult to prove claim in itself e the minibusetaxi industry provides other employment opportunities not directly dependent on the transport system (i.e. maintenance services) that will be directly impacted. There are other obstacles to wholesale system transformation based on full specification BRT, not directly related to the ‘paratransit’ sector. These relate to low anticipated passenger volumes, particularly in the first phase of the system, and the related need for subsidies to cover operating deficits and exposure to financial risk at the municipal level. The choice for the first phase of the system (be it Phase I or Phase 1A) was based on corridor characteristics. The main problems to be addressed were the high traffic congestion levels and the lack of suitable public transport alternatives. Other characteristics of the West Coast corridor were the fewer number of minibusetaxi operators and a greater catchment of choice users. Phase I was thus presented as a basis upon which the viability of BRT operations would be tested. However, the anticipated passenger demand, in the short term at least, is low when compared with international BRT systems in operation. According to official estimations, the forecast maximum peak hour passenger load for the West Coast corridor when full services are launched is 1307 passengers/hour/direction (Fortune, 2011), with other peak link estimations reaching 2076 passengers/hour/direction when a 10% modal shift from private vehicles is included (City of Cape Town, 2010). These loads are within the range conventionally considered appropriate for regular buses12 (a maximum capacity of approximately 8000 passengers/hour/direction (Lleras & Pereira, 2009; Vuchic, 2007; Zhang, 2009)). Different BRT systems’ loads are shown in Table 3; the values are well above those of Cape Town’s first phase corridor. Most developing world cities that have implemented BRT systems have very different operating environments when compared to South African cities: in general, urban densities are higher. Urban structures are also very different and can prove to be a major issue in the viability of BRTlike systems. A less expensive mode might be better suited to lower volume corridors in Cape Town. Conventional buses with road space prioritisation and an enhanced quality of service could provide similar benefits at less cost. Passenger volumes in the initial operations could create difficulties in the implementation of further phases as the sustainability and viability of the initial test might well be questioned by decision-makers. Passenger volume estimations along other corridors are projected to be higher than those of Phase 1A. For the Klipfontein Road corridor for instance, a maximum peak hour demand of 8760 passengers per direction was estimated (Wilkinson, Behrens, & Palmer, 2006). However, in this case, investment in duplicating rail- and road-based systems is problematic as this implies a duplication of operating subsidies along a single transport corridor. Low passenger volumes can also create important complications in terms of financing the system’s operations. The initial conception of subsidy-free systems seems highly unlikely, and indeed the operating subsidies needed might be higher than current levels.

12 ‘Regular buses’ are defined in this paper as bus services using conventional vehicles operating in mixed traffic.

125

Table 3 Peak loads in selected BRT corridors by city. Bogota Santiago Curitiba Mexico C. Beijing Jakarta Colombia Chile Brazil Mexico China Indonesia Maximum load 45,000 (pax/h/dir)

22,000

13,000

9000

8000

3600

Source: Hidalgo & Carrigan, 2010.

Initial high-level calculations suggest that significant subsidies will be needed for the BRT network (Palmer Development Group, 2011), as well as for compensation of minibusetaxi operators (McLachlan, 2010). The introduction of the new ‘formal’ system will require higher operating subsidies than the current subsidies given to transport operators. It is possible that, for six South African cities,13 the required operating subsidies will almost double from R5.3 billion per annum to R10.3 billion per annum (Palmer Development Group, 2011). In this same calculation, Cape Town’s subsidies would increase from approximately R2.1 billion per annum to R3.5 billion per annum. Because the projected system will be based on gross cost contracting, risks will be taken by the City and not by operators. This situation could create obstacles when introducing later phases of the system, as financially the City will take on significant risk, which it has never previously had to bear. In summary, the obstacles confronting total implementation of the BRT network plan appear to be significant. Continuing resistance from ‘paratransit’ operators to the imposition of IRPTN operations in other corridors, and low passenger volumes and the high level of required subsidies and associated risks, present significant challenges. Even if the whole proposed network is built and the transformation of the city’s public transport system is achieved, it will take considerably longer than initially envisioned. A likely outcome for Cape Town is that the city will depend for a long time on a ‘hybrid’ system that combines both ‘formal’ and ‘paratransit’ modes. Complex questions regarding how such ‘hybridity’ should be managed and regulated need to be addressed. The main objective of this paper is then to analyse urban transport systems that are experiencing similar transformational processes, and to draw lessons that could eventually be applied in Cape Town. 3. Learning from experience with ‘hybrid’ systems elsewhere 3.1. Rationale for selection of case studies ‘Hybrid’ public transport systems are a common feature of developing world cities. Even South American cities that have been described as exemplary examples of BRT implementation, have typically been unable to eradicate or formalise their ‘paratransit’ operations.14 These cities are still highly dependent upon ‘paratransit’ services, in addition to the existing and new ‘formal’ services. It is thus pertinent for Cape Town, as well as for other

13 The six considered cities were: Cape Town, eThekwini, Johannesburg, Ekurhuleni, Nelson Mandela Bay Metro and Tshwane. 14 Transformational processes in Curitiba, Bogota and to a lesser extent Quito, are considered exemplary South American cases. Bogota and Quito have not completely substituted or formalised their ‘paratransit’ services; only Curitiba has done so. However, this transformation was achieved under fairly unique circumstances. Firstly, public transport system improvement was a key focus of an unusually stable and sustained city planning process. Secondly, the formalisation of ‘paratransit’ operators was initiated early in the city’s development and growth, and thus it was done while the city was being built as opposed to when the city was already built. Thirdly, BRT system implementation came after ‘paratransit’ services were formalised. Curitiba’s outcome was achieved over a long period of time and it was supported by an urban plan with an important focus on the city’s public transport system.

126

P. Salazar Ferro et al. / Research in Transportation Economics 39 (2013) 121e132

South African cities, to study transformational experiences in international cities that can provide lessons on how to manage a ‘hybrid’ public transport system composed of both ‘paratransit’ and ‘formal’ operations. Experiences were classified according to the role of ‘paratransit’ services in the projected final public transport system and their role during the planning process. Only developing world cities that have ‘hybrid’ urban transport systems were considered, as the main objective was to draw lessons on how to manage the relationship between ‘paratransit’ services and ‘formal’ services. Three groups of experiences were identified: cities depending on a de facto ‘hybrid’ system; cities exhibiting a modified de facto ‘hybrid’ system; and cities serviced by a de jure ‘hybrid’ system. De facto ‘hybrid’ systems describe cities that have planned and initiated major public transport transformation processes. These proposals usually start with the implementation (phased or not) of a new ‘formal’ transport system along high demand or profitable transport corridors, while advocating the removal or substitution of ‘paratransit’ services. However, even in successful experiences, ‘paratransit’ operators retain a large e if not the largest e share of the commuter modal split in the city and they continue to operate in a poorly regulated environment with no planned integration with the new ‘formal’ system. A second set of experiences concerns cities that were prepared to roll out new ‘formal’ systems and replace incumbent ‘paratransit’ operations, but were forced to reconsider their initial proposals and ended up with a different project than first imagined. ‘Paratransit’ services are recognised and better accommodated through modifications to the initial scheme. This group of experiences is referred to as modified de facto ‘hybrid’ systems. A third set of experiences is the de jure ‘hybrid’ public transport systems. In this group, cities have recognised early in the planning process that a ‘hybrid’ system is inevitable. Their main objective is then to formalise and regulate the current transport system (usually composed of a majority of ‘paratransit’ operations) via fleet renewals and regulatory mechanisms to control the issuing of route or zonal licenses. They thus accept the importance of the ‘paratransit’ sector in the provision of transport in dynamic and growing cities. Regulatory frameworks are created according to this approach. Secondary sources were utilised in order to classify different transformational processes into one of the three groups. For each group, a short list of cities was identified; and case studies were selected. Available information and the possibility to contact researchers that have studied the urban public transport systems in these cities were important criteria in the final choice of case studies. This document focuses mainly on the de facto and de jure ‘hybrid’ transport systems’ case studies. As discussed before, de facto ‘hybrid’ transport systems lack the recognition of a system composed of ‘paratransit’ and ‘formal’ services. The selected case for this document is Bogota, in Colombia. This city has a worldrenowned new ‘formal’ public transport system in Transmilenio, but ‘paratransit’ services are still responsible for the majority of motorised public transport trips. There is limited planning for ‘paratransit’ services and, as a result, they continue to compete, under different regulations, with Transmilenio. In contrast to this type of experience, de jure ‘hybrid’ transport systems envision ‘paratransit’ services as the starting point of a transformation process. This is the case of Accra, in Ghana, where a phased project of formalising ‘paratransit’ operations is currently being implemented. The planning framework was developed around the role of ‘paratransit’ services, and the introduction of new ‘formal’ systems is a later stage in the process. These two cases thus present two

extremes in a spectrum of approaches to public transport transformation. Cape Town initially resembled the Bogota experience as it was expected that ‘paratransit’ services would be substituted by new ‘formal’ services, and the planning and regulatory framework for ‘paratransit’ operations would be interim in nature. However, consideration is currently being given to modest recognition of, and a role for, branded ‘paratransit’ feeder services in the first phase. Table 4 presents system characteristics and performance indicators that illustrate the need for public transport system transformation in the selected case studies. In the case of Bogota, the main justification for the implementation of Transmilenio was an overall reduction of travel time across the entire mobility system of the city (Lleras, 2005). This objective was achieved as public transport travel times decreased by 32% and mean speed increased by approximately 12% (Echeverry, Ibáñez, & Moya, 2005). However, important differences still exist between corridors where the BRT has been implemented (where speeds have increased) and corridors where BRT has not been implemented (where speeds have decreased (Echeverry et al., 2005)). In the case of Accra, the need for public transport system transformation is linked mainly to congestion problems associated with the current unregulated state of the system, and the need for structural changes to the Table 4 Key performance indicators.

Road network average speed (peak periods) Modal split Paratransit services Planned services Vehicles (road-based) Paratransit services Planned services Fare per trip (road-based) Paratransit services Planned services Road network average speed (peak periods) Modal split Paratransit services Planned services Vehicles (road-based) Paratransit services Planned services Fare per trip (road-based) Paratransit services Planned services Road network average speed (peak and off-peak periods) Modal split Paratransit services Planned services Paratransit services Planned services Vehicles (road-based) Paratransit services Planned services Paratransit services Planned services Fare per trip (road-based) Paratransit services Planned services Paratransit services Planned services

Cape Town

Year

Source

km/h

35

2009

Roux, Del Mistro, & Mfinanga, in press

% %

29 71

2009

Roux et al., in press

e e

7467 1160

2009

Roux et al., in press

US$ US$

(variable) 2.05

2005

Roux et al., in press

Accra 18

Year 2010

Source UATP & UITP, 2010

% %

84 16

2008

Kumar & Barrett, 2008

e e

6000 600

2008

Kumar & Barrett, 2008

2008

Kumar & Barrett, 2008

km/h

US$ US$ km/h

0.06e0.39 Bogota 27 32

% % % %

80 20 100 0

e

20,500

e

1190

e e

22,000 0

US$ US$ US$ US$

e

Year Source 2000 Echeverry et al., 2005 2005 2010 Alcaldia de Bogota, 2011 2000

2005 Ardila Gómez, 2007, pp. 9e15 2010 Hidalgo & Carrigan, 2010 2000 Ardila Gómez, 2005 Authors

0.84 2012 Authors 1.12 0.40 2000 Echeverry et al., 2005 Authors

P. Salazar Ferro et al. / Research in Transportation Economics 39 (2013) 121e132

127

Fig. 2. Urban operating environment in Bogota. Transmilenio routes and existing public transport routes. Source: Alcaldía Mayor de Bogotá, 2011. Public transport origins Public transport destinations. Source: Duarte Guterman & Cal y Mayor, 2006.

organisation and regulation of the paratransit industry (Finn, Abeiku-Arthur, & Gyamera, 2009). The analysis of modified de facto cases will be presented in an introductory manner. Possible relationships between ‘paratransit’ and ‘formal’ services will be presented and the basis for further research will be established. Ultimately, this group of cities could prove to be the more relevant for the management of Cape Town’s ‘hybrid’ urban public transport system. 3.2. The Bogota case In 2000, Bogota began implementing a full-specification BRT system along selected transport corridors. The Transmilenio system’s implementation called for a phased approach that would ultimately cover the entire city by 2030, substituting ‘paratransit’ operators by including them in newly formed BRT operating companies. The plan thus called for a general overhaul of the existing private sector-provided and chaotic urban public transport system. A plan to modernise and upgrade the existing transport system was undoubtedly required. Before the implementation of the

Transmilenio system, Bogota’s traffic conditions were poor and the quality of service offered by the public transport system did little to ameliorate a general urban crisis (Salazar Ferro, 2008). In 1998, despite relatively low levels of motorisation (approximately 142 vehicles per 1000 inhabitants15), congestion and accessibility problems were all too common due, in part, to a deficient public transport system composed entirely of private ‘paratransit’ bus companies (Salazar Ferro, 2008). Urban mobility conditions were most problematic in peripheral areas of Bogota, where high-density middle- to low-income residential neighbourhoods were located. The newly introduced BRT system benefited from Bogota’s unique operating environment to achieve high levels of service. First, urban population densities in Bogota are high: the overall density is 175 inhab/ha, but the density in the built-up area is 214 inhab/ha (Alcaldía Mayor de Bogotá, 2011). When compared to the standard urban densities of Latin America (90 inhab/ha) or Africa

15 The source states that there were 650,000 vehicles for 6,000,000 inhabitants in the city (Salazar Ferro, 2008). This information was used to calculate the car ratio per 1000 inhabitants.

128

P. Salazar Ferro et al. / Research in Transportation Economics 39 (2013) 121e132

(102 inhab/ha) (Lupano & Sánchez, 2009),16 Bogota is a very densely populated city. Furthermore, the areas exhibiting the higher population densities are located away from the employment and activity areas, in peripheral zones occupied by middle- and lowincome groups (Alcaldía Mayor de Bogotá, 2011). A second important feature of Bogota is its urban structure: the major employment zone is linear in shape and it is accessed by important northesouth roads where many transport routes converge (Fig. 2). The previous condition creates heavy transport demands along relatively low distances. The first phases of the Transmilenio system then focused on linking different residential areas in the periphery of the city passing through the main activity corridor, benefiting from the favourable urban structure and high passenger demands. Transmilenio is renowned for having achieved a remarkable peak hour ridership of 45,000 passengers/hour/direction (CAF, 2010; Duarte Guterman & Cal y Mayor, 2006; Hidalgo & Carrigan, 2010). While this achievement has been paradigm defining in terms of the proven capacity of BRT systems, it is unlikely that Bogota’s operating conditions can easily be reproduced in other cities or even in other corridors of the city. This peak hour ridership was measured where many BRT routes converge in the busiest part of the city, but single corridor ridership is more likely to be in the 20,000e30,000 passenger/hour/direction range.17 Future phases of the BRT implementation plan will not benefit from such high passenger demands as they are planned away from the main activity zones.18 In 2011, after ten years of existence and the implementation of the system in the more profitable transport corridors of Bogota, Transmilenio officially serves less than 20% of public transport trips in the city, while the remaining 80% or more is served by ‘paratransit’ operators (Alcaldía Mayor de Bogotá, 2011). Substitution of ‘paratransit’ operators has not been achieved. Recent documents suggest that, after an initial period of success, city traffic conditions and ridership in Transmilenio are declining (Gilbert, 2008). Another concern is the number of public transport vehicles in Bogota: one of the main objectives of the new system was to reduce the number of ‘paratransit’ vehicles operating in the city. Phase I19 called for the newly formed bus companies to scrap 2.7 old buses for every BRT bus that was introduced, and Phase II called for 7.7 old buses for every BRT bus introduced20. As a result, in 2006, it was expected that only 10,000 paratransit buses would still be operating, but the number of paratransit vehicles for that year was estimated at 20,847 vehicles, divided into buses, midi-buses and minibuses types (Gilbert, 2008). In 2004, official documents estimated that the total monthly number of operating licensed ‘paratransit’ vehicles was 17,118 vehicles: of which 6588 were conventional buses, 6356 were midi-buses and 4174 were minibuses (Duarte Guterman & Cal y Mayor, 2006). These data suggest that the number of ‘paratransit’ vehicles has not been reduced and that it is likely that the total number of ‘paratransit’ vehicles

16

Cape Town’s density is 39 inhab/ha (Wilkinson, 2000). Using the expected daily demand information presented in the Bogota mobility plan (Duarte Guterman & Cal y Mayor, 2006), peak demand information was estimated to be 20%e25% of the daily value. 18 The corridors included in future phases of BRT construction could change in the coming years as a metro system has been proposed. The introduction of such a system will undoubtedly require significant modifications to the BRT plan. 19 The Transmilenio project is composed of eight phases. Currently, the city is building Phase III. However, the current Mayor of Bogota, Samuel Moreno Rojas, was elected under the promise that the city would build a metro system and thus plans for the complete roll out of the BRT system have since been significantly modified. 20 In roads where BRT lanes were constructed, ‘paratransit’ services were forced out of the road. As a result, along Transmilenio corridors, the system has no direct competition. 17

operating in Bogota has indeed increased after the introduction of Transmilenio. The increasing number of ‘paratransit’ buses in Bogota has resulted in a decreasing vehicle ridership rate. For efficient utilisation, a conventional bus should ideally carry between 800 and 1000 passengers/day: in Bogota, the mean ridership rate is approximately 250 passengers/day (Ardila Gómez, 2005). It has been estimated that there is an excess of more than 7500 buses in the city (Ardila Gómez, 2005). This problematic condition is attributed to deficient market entry control by transport authorities (Duarte Guterman & Cal y Mayor, 2006), and an outdated faresetting policy that provides no disincentive for more inefficient vehicle use (Ardila Gómez, 2007, pp. 9e15). Under the circumstances described above, private ‘paratransit’ bus companies have limited interest in being included in transformational processes. In contrast to South African examples, in Bogota, the process of inclusion of the ‘paratransit’ sector was initiated by the bus companies. However these companies generally do not own vehicles as they are mainly concerned with generating income by affiliating individual buses (Ardila Gómez, 2005). Individual operators (and owners of ‘paratransit’ vehicles) were largely left out of the BRT implementation process. The result of Bogota’s approach is that the city is supported by a dual system where ‘formal’ services will most likely experience difficulties in implementing new corridors, and ‘paratransit’ services will continue to be profitable and thus difficult to include in transformational initiatives. Currently, the City is engaged in negotiations e independent of BRT implementation e with incumbent ‘paratransit’ operators to ‘formalise’ their operations. These negotiations were made necessary well after the implementation of the BRT system, suggesting a flaw in the initial model. 3.3. The Accra case Accra’s approach to systemic public transport transformation did not begin with ‘paratransit’ services being the main focus. In 2005, a report on the identification and prefeasibility of BRT corridors was published. The document, produced by DHV-The Netherlands in association with the Municipal Development Collaborative (MDC), evaluated design options and prepared cost estimates for a proposed implementation of a number of BRT corridors in the city (DHV-The Netherlands, 2005). However, recent developments show that the main focus of the transformational process has shifted to ‘paratransit’ services that will ultimately maintain their role as operators in the public transport system. As in the case of most developing world cities, Accra’s general mobility is highly dependent on ‘paratransit’ operations. The system exhibited important problems that needed e and still need e to be addressed: The urban transport problem is the expression of a stressed system which is typified by the absence of alternative transport like rail, poor quality public transport, low tech urban roads, the surge in on-street hawking of goods and services and the either weak and/ or poorly enforced urban transport regulations. This system claims about 1 600 lives and causes over 10 000 injuries on an annual basis. Source: Obeng-Odoom, 2009, page 50. The main component of the public transport system is ‘tro-tros’; a self-regulating sector composed of minibuses responsible for carrying 56% of daily passengers in the Greater Accra Metropolitan Area (GAMA) (DHV-The Netherlands, 2005). Typically these vehicles are old and they are not regularly maintained (DHV-The Netherlands, 2005). They appeared as a result of the collapse of a public transport company, and gradually increased their market share.

P. Salazar Ferro et al. / Research in Transportation Economics 39 (2013) 121e132

129

Fig. 3. 2004 and 2013 AM peak hour passenger demand in Accra (bus and ‘tro-tro’). 2004, current demand situation 2013, demand projections. Source: DHV-The Netherlands, 2005.

‘Tro-tros’ serve major activity nodes in the city, but they also provide services in recently developed areas where other transportation options are lacking (Finn, 2008). In a rapidly growing city such as Accra, their role becomes ever more important as they benefit from their natural flexibility and rapid response to serve new developments where ‘formal’ services have not yet been introduced. Furthermore, the emerging land use development pattern of the metropolitan area is one dominated by sprawl, based on low-rise single family detached units (Grant & Yankson, 2003). Travel distances have increased, as have public transport fares, thus

affecting the poorest sectors of society by excluding them from work opportunities (Finn et al., 2009). The low-density areas created by this type of development generate an operating environment where smaller buses can be the appropriate business model for the majority of the travel demand (Finn et al., 2009). Hence, the ‘tro-tro’ sector flourishes while larger ‘formal’ buses (requiring operating subsidies) would have difficulties matching the low passenger volume scenarios with appropriate services. Data presented in Fig. 3 for bus and ‘tro-tro’ demand, shows that currently only a few transport corridors have peak hour demands in

130

P. Salazar Ferro et al. / Research in Transportation Economics 39 (2013) 121e132

the 5000e7500 passengers/hour/direction range. Most are radial corridors entering the central area, and corridor sections nearing the ring road are those where demand is higher. In other corridors peak hour demands are lower than 5000 passengers/hour/direction. These corridors are located away from the CBD, in the urban peripheries, and in the inner roads of the central area. Projected peak hour demands for 2013 maintain a distribution across routes similar to 2004. For bus and ‘tro-tro’ demand, projected peak hour data shows that no corridors will exceed 7500 passengers/hour/ direction. Projected demand for areas away from the central area stays below 5000 passengers/hour/direction during the AM peak hour. Full-specification BRT operations along these corridors are not essential as enhanced bus services could cope with existing and future demands. A key point in the Accra transformation project was the early recognition that ‘paratransit’ services, in the form of ‘tro-tros’, have the potential to be significant contributors in the future urban public transport system. Acknowledging some structural weaknesses, most notably poor service quality and marginal profitability, ‘tro-tros’ are nevertheless relatively well organised, well established and well understood by the population in Accra and other cities in Ghana (Finn et al., 2009). Therefore, the Ghanaian initiative seeks an improvement of the ‘paratransit’ sector, avoiding processes that can marginalise or diminish it (Finn et al., 2009). The projected outcome is an efficient and modernised public transport system meeting the demands and expectations of the general population, and matched to institutional capacity and public finance limitations. As the project is still in its first stages,21 results and analyses are not yet available, but the study of the proposed framework provides insight into how a ‘paratransit’ sector might be gradually upgraded. The gradual transformation process of the public transport system is composed of three main stages. The first of those stages tackles the ‘legalisation’ of ‘paratransit’ operations by requiring operators to register as service providers and to have a valid license/permit. Implicitly when registering, ‘paratransit’ operators accept that they will ultimately change from a self-regulating enterprise to a (city) publicly regulated one. During this first phase, the city should acquire valuable data on current ‘paratransit’ operations, and it is expected that this will help in addressing overtrading problems and route competition disputes (Finn et al., 2009). The second phase of the initiative is the issuing of type ‘A’ permits to ‘paratransit’ operators. Operators are encouraged to upgrade their performance by gradually upgrading their quality standards. Type ‘A’ permits, issued for twelve months and renewable subject to performance, will be needed in order to be able to stay in the system; meaning that operators who do not comply with minimum requirements could be forced out of the public transport system (Finn et al., 2009). The proposed framework establishes flexibility to increase the number of vehicles on a route or to create new routes where no transport options are available. The final phase involving issuing of type ‘B’ permits,22 concerns ‘higher’ services (Finn et al., 2009) that include investments in (new) higher capacity vehicles and the consequent improvement of operational capacity. Type ‘B’ permits will have three-year validity, and minimum performance standards will require

21 The transformational process in Accra is currently in the first of three main phases. The second phase is expected to start in 2011. 22 Type ‘B’ permits will not supplant type ‘A’ permits. It is expected that routes offering type ‘B’ services will emerge from type ‘A’ routes and from operators involved in the transformation process. It is also expected that Metro Mass Transit services will be type ‘B’ services (Finn, 2011, pers comm).

services to be either scheduled or frequency controlled (Finn et al., 2009). The approach developed in the Accra case is consistent with the city’s urban growth, form and operating environment. The metropolitan area has relatively low densities23 and rural-like urban edges where high capacity transport systems are inappropriate. Furthermore, because of a lack of space in the central area, many new companies are located in the outskirts of the CBD and have created three identifiable nodes of corporate headquarter activities (Grant & Yankson, 2003). This in turn has created new demands in terms of transport destinations that can be best served by flexible and demand responsive systems such as ‘trotros’. The transformational project focuses on the gradual improvement of ‘paratransit’ operations. Coupled with the development of high capacity services along selected transport corridors, it is expected that the initiative will restructure the city’s public transport system and ‘formalise’ ‘paratransit’ services. This approach does not preclude the introduction of BRT systems, as they are considered as a possible step in later phases of the transformational process. 3.4. Modified de facto cases: possibilities for integration In between examples such the Bogota case and the Accra case, many developing world cities have implemented transformative processes where ‘paratransit’ operators have acquired officially sanctioned roles that were not intended for them initially. Cities undergoing these systemic transformations illustrate that the implementation of catalyst BRT projects does not necessarily imply the eradication of ‘paratransit’ services, and that ‘paratransit’ and ‘formal’ services can complement each other e as opposed to undermining each other’s performance. Cities in this category are not necessarily exemplary cases, but they warrant further research as they may provide valuable information on possibilities with respect to the ‘paratransit’ e ‘formal’ service relationship. From understanding and analysing these relationships, it may be possible to draw lessons for the South African cases, and more specifically for Cape Town’s transformational process. Cities in this category can begin to define the role of ‘formal’ operators and the role of ‘paratransit’ operators. More importantly, these cities can shed light on possible integration and complementarity in public transport system transformation processes. Integration and complementarity are logically context dependent and particular situations (in terms of urban environment and regulatory frameworks) have an important influence in appropriate solutions. A review of secondary sources on modified de facto cases revealed four options with respect to managing the relationship between ‘formal’ and ‘paratransit’ services. Each one of these options proposes different roles for ‘paratransit’ operators within the hybrid public transport system, and requires varying levels of change. The presented options are described in turn, starting with the alternative with least service interaction, and finishing with the alternative with greatest service interaction.24 A first approach to consider is the parallel or separate corridors alternative. In the case of parallel corridors, by having ‘formal’ services on defined routes, it is possible to complement them with ‘paratransit’ services on parallel roads. This model could split the demand according to the respective capacities of ‘formal’ and

23 The general density of Accra went from 140 inhabitants/ha in 1990 to 80 inhabitants/ha in 2005 (Finn et al., 2009). 24 This ordering does not imply an evaluation of these alternatives.

P. Salazar Ferro et al. / Research in Transportation Economics 39 (2013) 121e132

131

‘paratransit’ services. In the case of separate corridors, it is generally accepted that higher capacities and longer distances are associated with ‘formal’ services, while ‘paratransit’ operations usually rely on smaller vehicles (and reduced capacities). Such types of vehicles, generally minibuses and midi-buses in African cases, are not suited to high density e high demand corridors (Browning, 2001). The complementarity of the model is thus explained by an important distinction between the passenger volumes served by each type of operation on different corridors. The second possibility has two options with regard to a shared public transport corridor. The first is to have an exclusive public transport lane where both ‘paratransit’ and ‘formal’ operators could operate away from general traffic. Some cities are operating this system in practice (e.g. Delhi, in India); while others are in the implementation phase (e.g. Port Elizabeth). The second option consists of an exclusive ‘formal’ public transport lane with ‘paratransit’ vehicles operating in parallel general traffic lanes. The third possibility for ‘formal’ and ‘paratransit’ integration is peak-lopping. Few cases have formally employed this option in the developing world and developed world alike. The model consists of contracting ‘paratransit’ to operate on ‘formal’ service routes during peak hours. The greater number of vehicles can be translated into higher frequencies and more system capacity. During off-peak hours, the lesser demand would be served by ‘formal’ operations only. This model permits a temporary complementarity between previously competing modes and can prove effective by reducing operating costs induced by the differences between peak hour and off-peak hour demands. Smaller ‘formal’ vehicle fleets can be used while maintaining an adequate performance throughout the day. Finally, a trunk-‘formal’ and feeder-‘paratransit’ model presents further possibilities for ‘hybrid’ transport system management. Higher capacity ‘formal’ services would occupy the main, longer distance transport corridors, and be fed by smaller capacity ‘paratransit’ vehicles at specific points in the network. The most important difference with the first approach identified above is the provision of interchange points between ‘formal’ and ‘paratransit’ systems where transfer between services is facilitated. A review of international experiences suggests that there are three main trunk and feeder arrangement options: reward schemes; feeder area licencing; and franchising or concessioning.

improvement of public transport systems (Sohail, Maunder, & Cavill, 2006). As discussed earlier, Cape Town might initially have been classified as a de facto ‘hybrid’ case, but recent developments suggest a possible shift towards modified de facto characteristics might be in the offing. Further case studies of this type of hybrid system may therefore provide valuable lessons. The extraction of lessons from modified de facto cases, however, needs to take into account differences in operating environments and in the cities’ national and local regulatory frameworks. Furthermore, differences in the structure and organisation of ‘paratransit’ systems need to be understood. The applicability of lessons will depend upon, amongst others, the compatibility of the solution with the operating environment. Different urban characteristics and urban structures are more or less suited to, for example, the trunk and feeder alternative. There is a need to clearly define the public transport problem to be addressed within a particular urban context before selecting and introducing a solution. Implementing catalytic BRT projects need not imply immediate substitution of ‘paratransit’ services. As presented in the analysis of the Bogota case, ‘paratransit’ operations have lost territorial space in the city, but at the same time have maintained their share of the public transport passenger market. This has resulted in degraded traffic conditions in certain areas of the city. Bogota may be an exemplary case in terms of BRT implementation, but it is not an exemplary with regard to ‘formal’ e ‘paratransit’ system integration. Early recognition of the importance of ‘paratransit’ services, and their adaptability to changing urban environments, is difficult to achieve because of political pressures to provide immediate and comprehensive relief to urban traffic problems, often favouring a particular mode. The Accra case presents an encouraging example of where such early recognition of the role of ‘paratransit’ services has been achieved. Its gradual and flexible approach to the ‘formalisation’ of ‘paratransit’ services does not, however, preclude the implementation of BRT on high capacity corridors. The proposed transformation of a deficient transport system is supported by a regulatory framework grounded in actual conditions. This framework clearly defines the role of ‘paratransit’ operators and, at the same time, it allows them to gradually evolve into better quality services and sustainable operations. The approach recognises the importance and viability of ‘tro-tros’ in Accra: minibuses are well suited to the prevailing low densities and increasing trip distances.

4. Discussion: drawing lessons for systemic public transport transformation projects

5. Conclusions

Many of the transformational processes documented in the literature could be classified as de facto systems, as typically BRT implementation is expected to replace ‘paratransit’ services. Using the analysis presented on Bogota, one of the first lessons drawn is that the complete substitution of ‘paratransit’ services is highly unlikely. For many South African cities, and most notably Cape Town, the likely outcome of a ‘hybrid’ public transport system, requires attention is given to regulatory frameworks aimed at complementary ‘formal’ and ‘paratransit’ services. Considering the existing policy dichotomy discussed earlier, without public transport policy reform, Cape Town will rely on regulatory frameworks that are poorly suited to actual public transport service conditions. Acknowledgement of a future system composed of ‘formal’ and ‘paratransit’ services is difficult under the current national regulatory framework. Elsewhere is has been concluded that a regulatory framework inclusive of both ‘formal’ and ‘paratransit’ services is an important precondition for the

The inhabitants of developing world cities are normally heavily dependent upon ‘paratransit’ services for their daily mobility. Most current initiatives aimed at overhauling public transport systems address valid deficiencies of ‘paratransit’ services, and propose a complete substitution of such operations. The result of this approach has generally been the appearance or endurance of a complex set of independent ‘formal’ and ‘paratransit’ operations, very often subject to parallel or disconnected regulatory frameworks. These ‘hybrid’ systems are a common feature of developing world cities. Policy recognition of the ‘paratransit’ sector as an important component of public transport systems should lead to a healthier ‘formal’ e ‘paratransit’ relationship. This, in turn, should lead to more appropriate frameworks on how to manage ‘hybrid’ systems and to achieve (urgent) desired improvements to currently deficient systems. Improved management of ‘hybrid’ transport systems does not exclude the implementation of BRT projects. Such initiatives need not eliminate ‘paratransit’ services. In an ideal ‘hybrid’

132

P. Salazar Ferro et al. / Research in Transportation Economics 39 (2013) 121e132

public transport system, BRT operations and incumbent ‘paratransit’ services are aimed at a common objective of creating the best possible urban transport system for a city, subject to the constraints and path dependencies presented by the local context. Acknowledgement The research presented in this paper was funded by the Volvo Research and Educational Foundations, and forms part of a broader research programme conducted by the African Centre of Excellence for Studies in Public and Non-motorised Transport (ACET, www. acet.uct.ac.za). Paul Browning and Brendan Finn provided comments on an earlier version of this paper. After the presentation of a first draft of this paper in a workshop session at the Thredbo 12 Conferences, revisions were made to reflect a further set of comments by Juan Carlos Muñoz (chair of the workshop). References Alcaldía Mayor de Bogotá e Secretaría de Planeación. (2011). Documento Técnico de Soporte e Modificación al plan de Ordenamiento territorial de Bogotá. Bogota, Colombia: Alcaldía Mayor de Bogotá. Ardila Gómez, A. (2005). La olla a presión del transporte público en Bogotá. Revista de Ingeniería, 21. Ardila Gómez, A. (2007). How public transportation’s past is haunting its future in Bogota, Colombia. Transportation Research Record no 2038. Browning, P. (2001). Wealth on wheels? The minibus-taxi, economic empowerment and the new passenger transport policy. In: 20th SATC Conference, 16e20 July. Pretoria, South Africa. CAF. (2010). Observatorio de movilidad urbana para América Latina. Bogota, Colombia: CAF e Corporación Andina de Fomento. City of Cape Town. (2005). Summary of 2004/5 current public transport record. Cape Town, South Africa: City of Cape Town. City of Cape Town. (2007). Public transport implementation framework: Integrated public transport network. Cape Town, South Africa: City of Cape Town. City of Cape Town. (2010). Business plan e Phase 1A of Cape Town’s My CiTi integrated rapid transport system. Cape Town, South Africa: City of Cape Town. City of Cape Town. (2011). Integrated rapid transit project e Progress report no. 14. Cape Town, South Africa: City of Cape Town. Clark, P., & Crous, W. (2002). Public transport in metropolitan Cape Town: past, present and future. Transport Reviews, 22(1), 77e101. DHV-The Netherlands. (2005). Bus rapid transit options identification and prefeasability study: Draft final report. Accra, Ghana: Department of Urban Roads. Duarte Guterman y Cía. Ltda., & Cal y Mayor y Asociados. (2006). Formulación del plan Maestro de movilidad para Bogotá D.C. Bogota. Colombia: Alcaldía Mayor de Bogotá. Echeverry, J. C., Ibáñez, A. M., & Moya, A. (2005). Una evaluación económica del sistema Transmilenio. Revista de Ingeniería, 21, 68e77. Finn, B. (2008). Market role and regulation of extensive urban minibus services as large bus service capacity is restored e case studies from Ghana, Georgia and Kazakhstan. Research in Transportation Economics, 22, 118e125. Finn, B. 2011. Personal communication. Finn, B., Abeiku-Arthur, B., & Gyamera, S. (2009). New regulatory framework for urban passenger transport in Ghanian cities. In: 11th Conference on competition and ownership in land passenger transport, 20e25 September. Delft, Netherlands.

Fortune, G. (2011). Cape Town integrated rapid transit project overview. Lecture presentation. University of Cape Town, Centre for Transport Studies, 31 March 2011. Gilbert, A. (2008). Bus rapid transit. Is Transmilenio a miracle cure. Transport Reviews, 28(4), 439e467. Grant, R., & Yankson, P. (2003). City profile: Accra. Cities, 20(1), 65e74. Grey, P. (2006). Regulating the informal. Exploring the possibilities for formalising the South African minibus taxi industry. Cape Town, South Africa: Masters dissertation e University of Cape Town. Hidalgo, D., & Carrigan, A. (2010). Modernizing public transport e Lessons learnt from major bus improvements in Latin America and Asia. Embarq e World Resources Institute. Kumar, A., & Barrett, F. (2008). Stuck in traffic: Urban transport in Africa. World Bank. Lleras, E. G. C. (2005). Transmilenio y el transport colectivo tradicional, una relación incierta. Revista de Ingeniería, 21, 84e93. Lleras, G., & Pereira, L. (2009). No recipe for success. Mobility, (July/September), 32e35. Lupano, J. A., & Sánchez, R. J. (2009). Políticas de movilidad urbana e infrastructura urbana de transporte. United Nations, Santiago, Chile: CEPAL. McLachlan, N. (2010). The introduction of bus rapid transit systems in South African cities. Participation of the minibus taxi industry e A model for sustainability or a recipe for failure. CODATU XIV, 25e27 October. Buenos Aires, Argentina. National Department of Transport. (2005). Key results of the national household travel survey. Pretoria, South Africa: Department of Transport. National Department of Transport. (2009). Technical transport planning guidelines for Comprehensive Integrated Transport Plans (CITPs) to be prepared by type 1 planning authorities. Pretoria, South Africa: National Department of Transport. Obeng-Odoom, F. (2009). The future of our cities. Cities, 26, 49e53. Palmer Development Group. (2011). Assessment of financial trends for metro public transport functions. Cape Town, South Africa. Roux, Y. E., Del Mistro, R. & Mfinanga, D. Comparative analysis of public transport systems in African cities. In: 2012 SATC Conference. Pretoria, South Africa, in press. Salazar Ferro, J. (2008). Bogota 1992e2005. The reversal of the crisis and the planning scheme. Megacities comparative study. Tokyo, Japan. Schalekamp, H., & Behrens, R. (2009). Engaging paratransit on public transport reform initiatives in South Africa: A critique of policy and an investigation of appropriate engagement approaches. In: 11th Conference on competition and ownership in land passenger transport, 20e25 September. Delft, Netherlands. Schalekamp, H., & Behrens, R. (2010). Engaging paratransit on public transport reform initiatives in South Africa: a critique of policy and an investigation of appropriate engagement approaches. Research in Transportation Economics, 29, 371e378. Schalekamp, H., & Behrens, R. (2011). Engaging the paratransit sector in Cape Town on public transport reform: progress, process and risks. In: Thredbo Conference, 11e15 September 2011. Durban, South Africa. Sohail, M., Maunder, D., & Cavill, S. (2006). Effective regulation for sustainable public transport in developing countries. Transport Policy, 13, 177e190. UATP, & UITP. (2010). Public transport in Sub-Saharan Africa. Major trends and case studies. Brussels, Belgium: Trans-Africa Consortium. Vuchic, V. R. (2007). Urban transit e Systems and technology. Hoboken, United States: John Wiley and Sons. Wilkinson, P. (2000). City profile: Cape Town. Cities, 17(3), 195e205. Wilkinson, P. (2008). Moving Ahead? La difficile transformation des systèmes de transports métropolitains au Cap. In A. Dubresson, & S. Jaglin (Eds.), Le Cap après l’apartheid e Gouvernance métropolitaine et changement urbain, Karthala. Wilkinson, P., Behrens, R., & Palmer, I. (2006). High level assessment of the viability of the Klipfontein corridor project. Cape Town, South Africa: Provincial Government of Western Cape. Zhang, M. (2009). Bus versus rail e meta-analysis of cost characteristics, carrying capacities, and land use impacts. Transportation Research Record, 2110, 87e95.