| dc.description.abstract | Background
Current longitudinal studies focus mainly on understanding how travel choices respond to changes in events that are within the control of individuals. Such events include changes in family structures, residential relocation, and acquisition of mobility tools. Because of this focus, little is known about the connection between travel behaviour and events outside the control of individuals. This knowledge gap is particularly prominent in Sub-Saharan African cities, where State actions generate conditions that significantly alter the available travel choices. These State actions include policies and strategies that are deliberately designed to influence travel choices as well as incidental outcomes of policies and strategies that are intended for other development goals.
Objectives
The current study aims to unearth the opportunities and threats that changes in the above conditions present for addressing the policy and infrastructure needs of pedal cycling in medium-sized Sub-Saharan African cities. Two specific objectives are pursued i) to trace the changes in mode choices at key points in the history of transport development in Kisumu between 1999 and 2014, and ii) to explain the drivers of these changes and their implications for addressing the needs of pedal cycling.
Data
The study employs retrospective categorical data obtained from Kisumu city in Kenya to realise the above objective. Kisumu is a typical medium-sized Sub-Saharan African city with about 400,000 inhabitants. The physical size of the city is approximately 297 square kilometres, although its urban footprint stretches only about 7 kilometres around the city centre. The remaining part of the city is under agricultural production. Like many other medium-sized Sub-Saharan African cities, Kisumu pursues various policies whose impact on travel choices remains scantily understood.
The data is obtained through interviews with 2,165 respondents drawn from 1,490 households. Its content includes the respondents’ places of residence, places of primary occupation, and the primary travel mode to this destination at each key date. A derived change in mode choice for each individual respondent and the main driver of this change is also included. The data relates to choices in the years 2004, 2009 and 2014. These dates are the culmination of various State efforts that supported the commercialisation of cycling for passenger transport, removal and reintroduction of taxes on pedal bicycles, reforms in public transport, emergence of motorcycle-taxi, and the evolving role of Kisumu as a university city.
Data processing and analysis
Changes in mode choices were determined by first calculating the proportion of respondents who used each mode in each base year and its corresponding turning point. The mode used in the year that immediately preceded the turning points was taken as the reference for the analysis of change. Thus the modal choices as at the year 2003 were the reference for the changes that occurred in 2004. Similarly, the choices as at the years 2008 and 2013 were the reference for the changes that happened in 2009 and 2014 respectively. Changes in mode choices were subsequently analysed by comparing the percentage share of each mode at every turning point against its corresponding share in the base year.
A multinomial logistic regression model was then developed to investigate the lag effect of the changing transport conditions on mode choices. Changes in mode choices at each turning point were modelled as the dependent variables while the corresponding reasons for the change were modelled as the predictors. The model works by estimating the probability associated with each value of the dependent variable before it identifies their predictors. The predictors of the changes are the conditions that were thought to emerge from the changing transport conditions in Kisumu.
Results and discussion
The results show a declining share of non-motorised modes amid a growing motorisation. A predominant revelation is the declining share of passenger cycling. Nonetheless, private cycling has re-emerged to stabilise the use of pedal bicycles in Kisumu. Supportive State policies are revealed to encourage cycling. However, this gain is diminished by the inferior social presentation by the State. Similarly, sympathetic State policies towards cycling are revealed to be incapable of achieving much in the long run if they are not validated by supportive road infrastructure. The study also reveals that the re-emergence of private cycling is an effort of its users to optimise emerging transport expenditure. This development is revealed to be a coping strategy of travellers whose demand for affordable transport can no longer be met by the new modes that are supported by the State. Changes in mode choices are also the result of changing preferences and generational change. The growth of private cycling by the youth is demonstrated by their preference for bicycles with gears, which they view to be better than the traditional ‘black mambas’.
Conclusions and recommendations
The paper demonstrates that analysing travel behaviour based on longitudinal data holds the potential to uncover the hidden drivers of changes in choices that would otherwise be lost from policy processes. The paper reveals the underlying accounts for the changes in mode choices over the years. It is revealed that the growth in motorcycling is a product of limited choices available to its users rather than their voluntary choice. The preoccupation of the State with facilitating pedal cycling and later motorcycling as a tool for income generation explains the diminishing share of pedal bicycles. At the same time, the re-emergence of private pedal cycling as a coping strategy against limited transport service remains invisible and unsupported by the State. The study concludes that transport policies and infrastructure development efforts must take cognizance of various State actions with a view to enhancing their positive opportunities while checking the developments that might hinder cycling. Examining the changes in mode choices on a yearly basis could greatly improve the detection of the drivers of the observed changes. | en_US |
