Sage Journals: Discover world-class research

Abstract

Vehicle ownership has increased tremendously in Nigeria for the past decades. This study determined factors that influence vehicle ownership intending to develop a more appropriate basis for forecasting vehicle ownership in the country. Multiple linear regression technique was used to identify the factors that influence vehicle ownership using national data from secondary sources. The results showed that four socioeconomic factors, namely, gross domestic product, per capita income, fuel price, and literacy level, as well as one physical factor, namely, stock of public transport vehicles, have significant effects on vehicle ownership at .01 significance level. Vehicle ownership was 35.3 million in 2018. It will increase to 48.7, 66.2, and 76.1 million in 2030, 2040, and 2050, respectively. Consequently, the formulation of appropriate policy that will be useful for monitoring key parameters is germane for predicting vehicle ownership in the country.

Keywords

vehicle ownership socioeconomic factors physical factors determinants forecast

Introduction

Globally, vehicle ownership has tremendously increased. It grew from 500 million in 1986 to 1 billion marks in 2010, thus doubling within 24 years (Dargay, 2007). A study by Wards (2014) showed that global vehicle registration increased by 3.6% from 980 million units in 2009 to 1.015 billion in 2010. Although this increase is felt notably in developed countries, some developing economies also show tremendous growth in the vehicle population. The rate of vehicle growth in middle-income developing countries of Thailand, Indonesia, and Brazil is projected to add about half as many as what is obtainable in the United States between 2003 and 2050 (Dargay, 2007).

In Nigeria, vehicle ownership has witnessed a tremendous increase in the past decades. Statistics show that it grew by 693% from 1970 to 2010. A breakdown of the statistics for some states such as Lagos, Abuja, Enugu, Kaduna, Akwa Ibom, and Bauchi shows that vehicle ownership grew phenomenally by an average of 134% (National Bureau of Statistics, 2018). The rapid growth in vehicle ownership, however, threatens the sustainability of the existing road transport system. It led to severe traffic congestion, a high rate of road traffic accidents, air pollution, deterioration of the roadway system, and excessive fuel demand. This also resulted in severe fuel scarcity that has bedeviled the country in the past decades (Agbonkhese et al., 2013; Nwachukwu & Chike, 2011).

Three major efforts were made by the Government to address the problems associated with vehicle ownership. These include the following: first, discouragement of automobile dependency through the imposition of high tariffs on private vehicles and zero tariffs on mass transit buses; second, the 1977 vehicle restraint regulation which mandated private vehicles with odd and even number plates to ply the roadway system in Lagos during odd and even number days of the week, respectively. This was targeted toward the elimination of traffic congestion resulting from automobile dependency. The third was the expansion of road capacities, construction of interchanges, bridges, and network of roads, in major cities such as Lagos and Port Harcourt.

The efforts have not yielded the desired results because rapid growth in vehicle ownership and its associated problems persist even on a higher scale in the country. Available statistics show that Nigeria has the second-highest road traffic accident rate in the world. Fuel demand has increased to higher levels, resulting in severe fuel scarcity. The deterioration of the roadway system, air pollution, traffic congestion, and vehicle ownership are now on a higher scale. The failure of the past efforts indicates that the parameters that influence vehicle ownership have not been satisfactorily identified by the policymakers. This indicates the inadequacy of the existing vehicle ownership forecast or the assumption that forms the basis for policy on vehicle ownership

Several vehicle ownership models based on aggregate and disaggregate data have been developed and applied in estimating and forecasting vehicle ownership. The applications of the models are rare in Nigeria and other African countries despite the continent’s unprecedented growth in vehicle ownership. Consequently, this study aims to model vehicle ownership in Nigeria. The model is intended to provide an accurate forecast that should form the base for formulating appropriate policy on vehicle ownership in Nigeria.

Literature Review

A plethora of model types have been applied in modeling and forecasting vehicle ownership around the globe. The two main aggregate time series model types that have been used to estimate and forecast vehicle ownership are the Gompertz function and linear regression model. Dargay and Gately (1999) used the Gompertz model to estimate the income effect on vehicle ownership worldwide from 1960 to 2015. The result indicates that there will be a convergence of vehicle levels, for most Organisation for Economic Co-operation and Development (OECD) countries, close to saturation levels over the next two decades. The United States will experience a relative increase in vehicle ownership, whereas vehicle ownership in low-income countries will double twice as rapidly as per capita income.

Similarly, Dargay et al. (2007) used the Gompertz model to determine factors that influence vehicle ownership worldwide up to 2030. The results indicate that income growth significantly influences vehicle ownership. The world vehicle ownership will increase by 2.5 times from 300 million in 2002 to 2 billion vehicles in 2030. Half of the world vehicles will be found in non-OECD countries during the period. In the same vein, Wu et al. (2014) applied the Gompertz model to forecast China’s vehicle ownership up to 2050. The results show that vehicle ownership will grow rapidly in China by three folds from 2012 to 2030. Subsequently, it will experience a gradual annual decline up to 2050, thus depicting an S-shaped curve. The inflection point of the increasing curve in vehicle ownership appeared around the year 2030, which shows an annual growth from 6.13% to 9.50% and declines to 0.45% in 2050. This is an indication that China is in the primary stage at Gompertz functions (i.e., vehicle ownership has not reached the saturation point). Furthermore, gross domestic product (GDP) was found to be the main driver of China’s vehicle ownership.

Also, Wang et al. (2007) used the Gompertz model to predict China’s vehicle ownership through to 2050. The results show that China will have more vehicles than the United States by 2030. It could have the largest number of vehicles in the world by 2035. The vehicle population will reach 486 to 662 million by 2050. This will pose enormous challenges to urban infrastructure, resources, and the environment. Lu et al. (2017) applied a new improved stochastic Gompertz delusion process to model China’s vehicle ownership up to 2025. The results indicate that vehicle ownership per 1,000 people and the total vehicle stock will be over 250 units and 350 million units in the year 2025, respectively. They suggest that the model is appropriate for forecasting vehicle ownership in the short run.

Huang (2011) used linear regression to forecast China’s car ownership. The results suggest that GDP, savings deposit, and highway mileage are the determinants of car ownership. The total car ownership will reach 66.84 million in 2015, resulting in 2.0 million barrels per day of oil consumption. Song and Wang (2017) applied Poisson and corrected Poisson regression models to measure vehicle ownership in the three counties of South Florida, USA. The results indicate that vehicle ownership is influenced by the proximity to schools, household income level, the number of drivers, net house density, housing tenure, and the number of workers. Also, there exist high levels of vehicle ownership among locally spatial clusters of the households, whereas it was statistically random for the globe. The outcome of the study is capable of enhancing measures toward environmentally friendly communities.

Several studies applied various types of discrete choice models to estimate and forecast vehicle ownership. Wong (2013) adopted a discrete choice approach (multinomial logit model) to analyze car ownership in Macao using disaggregate household data. The results indicate that the number of household members, number of children, household income, and population density are the underlying factors that influence vehicle ownership.

Zegras and Hannan (2012) used the multinomial logit choice model to examine household vehicle ownership decisions in Santiago, Chile. The results show that household preference for vehicle ownership changes over time and responds significantly to demographic socioeconomic and land-use factors.

Tsang et al. (2011) used the multinational logit choice model to forecast car ownership in the Sydney area. The study found that income, driving license, employment, accessibility, location, and household structure are the drivers of car ownership. Comparing the results with the previous studies, they found that holding the driving license has become a more important factor in vehicle ownership than income.

Tam and Lam (1999) used the multinational logit choice model to predict Hong Kong’s car ownership. They found that the household income, employment, vehicle license fee, home-end parking fee, and usage cost were sensitive to car ownership.

Matas and Raymond (2008) employed variations of both discrete choice models, namely, multinational logit choice model (ordered response mechanism) and ordered probit model (unordered response mechanism), to analyze factors that determine vehicle ownership in Spain. The results show that vehicle ownership responds positively to income, quality of public transportation, employment, and age of the population. There will be generational effects that will cause an immediate increase in vehicle ownership. The effects will decline progressively and disappear around the year 2020. Moreover, they found that both models are indistinguishable based on performance.

Ding et al. (2016) used a multilevel mixed ordered probit model to examine the heterogeneous effect of the physical development on household vehicle ownership levels in Washington-Baltimore Region, USA. The results show that the built environment accounted for 42.8% of spatial heterogeneity in household vehicle ownership. The outcome indicates that the model of this study is better than the traditional ordered probit model.

Gomez-Gelvez and Obando (2013) used multinomial logit and ordered logit models to forecast vehicle ownership in Bogota, Colombia. They found that income has the highest effect on vehicle ownership with an aggregate level elasticity of between 0.908 and 1.110. The impact of other variables, namely, household size, location, and population density, was low. This indicates that income is the greatest limitation of vehicle ownership.

Eakins (2015) applied two multinomial models to estimate household car ownership in Ireland across time. The result shows that car ownership is influenced by household expenditure and employment. This is an indication that the demand for car ownership increases with recovery in household expenditure and income.

Martha et al. (2017) used ordered logistic regression to examine the changing influence at various household car ownership levels in the Netherlands. The results show that the influence of household income, size, composition, gender, age, education, working status, and suburbanization levels on car ownership changed substantially between 1987 and 2014. The relative influences of household income and household size are substantial, contributing more than 60% of the total influence on household car ownership. The influence of household income on car ownership decreased over time from 38% in 1987 to 28% in 2014, whereas the influence of household size increases from 29% in 1987 to 35% in 2014. Household income and household size show a positive relationship with the number of cars owned by a household.

Ritter and Vance (2012) predicted vehicle ownership in Germany with the multinomial logit model. They found that vehicle ownership will increase moderately at 0.54% per annum until 2030. Furthermore, income, fuel price, driving license, and use features were the determinants of vehicle ownership. The outcome of the results can be used for comprehensive projections of emissions and congestion under alternative scenarios. Prabnask et al. (2011) used a multinomial logit approach to model household vehicle ownership in a medium-size Khon Kaen City, Thailand. The result showed that household income, the highest level of education, number of motorcycles, number of potential drivers, and number of car driving licenses were the determinants of vehicle ownership. The finding of this study is unique from large cities where the previous studies were undertaken in Thailand.

Pasra et al. (2018) used a multinomial logit model approach to examine car ownership of households in a suburban area in Makassar City, Indonesia. The study found that house types, family size, income, motorcycle ownership, number of family members working, and daily trip number are the determinants of households’ car ownership. The results may lead to the influence on the time valuation of the drivers and travel model choice of the households.

Joseph et al. (2017) applied the multinomial logit model in estimating the influence of household characteristics on car ownership in Akure, Nigeria. They found that the probability of owning a car increases with an increase in income and reduces with an increase in household members. Consequently, a 50% increase in income will lead to a 51% increase in vehicle ownership, whereas a decrease will result in a 37% deduction. On the contrary, a 50% increase or decrease in household size has a little influence of 0.8% increase or decrease in car ownership. The implication of this is that economic boom and recession have a significant effect on vehicle ownership.

The works of Kobos et al. (2003), Hirota (2007), and Le Vine et al. (2017) used static disaggregate car-type choice models to estimate vehicle ownership. Kobos et al. (2003) applied provincial-level logistic growth functions on simulation and scenario analysis of China’s vehicle ownership growth. The results indicate that growth in vehicle ownership is influenced by income and population growth rates. The passenger vehicles per 1,000 population will increase from 4.22 in 1995 to 54.33 in 2025. This will result in a 17-fold increase in oil demand and CO₂ emission.

Hirota (2007) used the JARI BAU (business-as-usual) model, a logistic approach, to estimate vehicle ownership and establish its determinant factors in Japan up to 2030. The model was chosen because it is extremely flexible and enables the easy explanations of the effects of explanatory variables on vehicle ownership. The results show that vehicle ownership was 57.28 million in 2010 and will rise to 62.59 million units in 2030. The socioeconomic factors of driver’s income and driver license holder population were found to have influenced vehicle ownership.

Le Vine et al. (2017) applied a binary logistic regression model to examine the determinants of household vehicle ownership in China. They found that there is an income effect on household car ownership. Also, there exists a negative association between vehicle ownership and living in rural areas, whereas the association was positive with living in urban areas. This is an indication that China’s vehicle ownership was accounted for by households living in urban areas.

In another vein, Dargay (2002), Cao and Huang (2013), and Yin et al. (2019) applied pseudo-panel-type models in their estimation of vehicle ownership. Dargay (2002) used the pseudo-panel approach to examine car ownership in rural and urban areas of the United Kingdom. He found that car ownership is more responsive to change in motoring costs for urban households than was the case for rural households. This implies that an increase in the cost of car transportation will portend a serious economic burden for rural households. Cao and Huang (2013) used panel data models (fixed-effects cluster data model and random-effects cluster data model) to examine city-level determinants of private car ownership in China. They found that larger cities and coastal regions experienced a higher level of car ownership and faster vehicle growth. The economic development variables of GDPs and disposable income per capita were found to be the most important determinants of vehicle ownership among different sizes of cities. Furthermore, the random-effects cluster model performs better than the fixed-effects model in the case of extra-large cities, whereas the fixed-effects model performed better than the random-effects model in the case of large, medium, and small cities. The socioeconomic (GDP, disposable income), spatial (rate of urbanization, size of built-up area), and transport-related (area of the roadway, stock of buses per 1,000 people, and stock of taxis) variables were found to be the determinants of vehicle ownership in different sizes of Chinese cities. Yin et al. (2019) employed a multi-group structural equation model to investigate the impact of the built environment on car use. They found that the environment has a significant effect on ownership and use. Consequently, urban planners and policymakers should use compact land-use strategies to reduce car ownership and use.

Yagi and Managi (2016) used the cohort model to examine demographic determinants of car ownership in Japan. They found that the elasticities of population and household size on car ownership are negative. This implies that a decrease in population and household size will result in an increase in car ownership. Furthermore, car ownership tends to be positively impacted by the concentration of population within prefectures as well as decrease and increase in relatively new and old cars, respectively. Berri (2009) used an age cohort period model to examine household car ownership behavior in seven countries characterized by different economic and cultural contexts. These are France, Italy, Japan, the Netherlands, Poland, the United Kingdom, and the United States. The study found that three main factors, namely, the history of car ownership development, the level of economic development, and population density, accounted for the differences in car ownership level between the countries. The United States where automobile diffusion started before World War II seems closer to saturation point than Western Europe, Japan, and Poland. Poland has the lowest motorization rate in the seven countries. Vehicle ownership rates are lower and sensitivity to income stronger in Italy. Finally, the more densely populated countries, namely, Japan and the Netherlands, recorded lower car ownership levels than the other industrialized countries.

Kermanshah and Ghazi (2001) used a two-level nested logit model to forecast vehicle ownership in Iran. The results indicate that vehicle ownership is influenced by household size, the number of workers, head of household’s age, occupation, life cycle, and life cycle stage. This is a demonstration that the model of this study is more appropriate than simple multinomial logit models in vehicle ownership forecast.

Ceylan et al. (2018) used multiple nonlinear regression analysis called the polynomial regression model to forecast vehicle ownership for the year 2035 in Turkey. They found that vehicle ownership will vary between 230 and 325 per 1,000 capita in the year 2035 depending on economic achievements, global oil prices, and national taxation policies. Also, the results show that vehicle ownership (car ownership) will be substantially lower in Turkey than that in the European Union countries despite that it has an increasing trend in the past two decades. Also, the number of employees, gasoline prices, car prices, and GDP per capita were found to be the determinants of car ownership.

Soltani (2017) applied the nested logit model in exploring the impact of socioeconomic status and urban form factors on the household car ownership in Iran. The results show that urban form (mixed-use) has a negative influence on household vehicle ownership, whereas socioeconomic characteristics (income and household size) showed a very strong positive effect. The outcome of this supports the argument in favor of the location of households in neighborhoods that facilitate their car ownership preferences.

Said (1992) used the generalized linear model approach to model household car ownership in Kuwait. Kuwait and two non-Kuwait households, namely, Arabs and Asians, were examined separately. The results show that area type and house type affected the car ownership rate of Kuwait households, whereas no such effect was found for Arab and Asian households. The model-estimated values were consistent with the observed car ownership levels. Also, household size, income, and number of adults in the household had a significant effect on Kuwait households’ car ownership. For Arab households, household size and income influenced their car ownership, whereas only income was found to be significant for Asians households.

Button et al. (1992) used an aggregate quasi-logistic approach to forecast car ownership in low-income countries. The results show that with the assumed GDP growth rate of 4%, the car ownership per 1,000 persons for Burkina Faso, Rwanda, Haiti, Pakistan, Cameroon, Gabon, Algeria, Mauritius, and Malaysia will reach 286, 449, 335, 161, 340, 440, 341, 209, and 284 in the year 2025, respectively. This implies that as low-income countries become more prosperous, there is an inevitable and rapid use in their car ownership rates. This reinforces temporal income effect as car ownership levels at any given level of income rise over time. Korkmaz and Akgungor (2018) estimated vehicle ownership in Turkey with the flower pollination algorithm model. They found that vehicle ownership will rise rapidly by 30% in 2025. Three variables were, namely, license number, domestic product per capita, and fuel prices, were found to account for the rapid increase in vehicle ownership.

Several studies applied a hybrid of model types to estimate and forecast vehicle ownership. Romilly et al. (2000) used a cointegration and general-to-specific approach to forecast car ownership in Britain. The results show that car ownership per 100 people will increase from 404 units in 1995 to 606 units in 2025, thus representing a 200% increment. They sounded a cautionary note on estimation results on grounds that the forecast may suffer from small sample bias. McCarthy and Wang (2017) used fixed-effects and Hausman-Taylor models to forecast China’s private car ownership. They found that for private car ownership it grew at an annual rate of 20%. The underlying factors that accounted for the growth were economic, infrastructure, spatial, and regulatory environments. They recommend the need to determine the impact of their study on traffic congestion, highway safety, air quality, and health.

Wu et al. (2014) applied the Gompertz function and computable general equilibrium models to forecast vehicle ownership in China. They found that the stock of vehicle will rise to 300, 455 and 465 million in 2050 for respective low-growth, medium-growth, and-high growth scenarios. The growth in vehicle ownership will increase beyond the Gompertz curve’s inflection point by 2020, but will not attain a saturation point before 2050. Furthermore, the study attributed the rapid growth in China’s vehicle ownership to per capita GDP. Das et al. (2010) used logistic and Gompertz function to model car ownership growth in Delhi, India. The results indicate that car ownership level will reach 14 and 17 cars per 100 persons by 2020–2021 under the saturation levels of 25 and 50, respectively.

Hao et al. (2011) used a hybrid model to estimate China’s vehicle ownership. The results show that the vehicle population would reach 184.8, 368.8, and 606.7 million by 2020, 2030, and 2050, respectively. The vehicle population in urban areas will account for 70.1%, 81.1%, and 86.1% of the total vehicle population in the respective years. Household income and vehicle prices were the driving factors accounting for the growth in vehicle ownership.

Ogut (2004) used three extrapolative models, namely, logistic, power growth, and Gompertz curves, to determine car ownership in Turkey. The results show that the logistic curve model overestimated car ownership between 1996 and 2002. The two other models, power growth and Gompertz curve, exhibited a similar forecast for 2020, thereby increasing the reliability of the forecast. Similarly, Serag (2014) applied three different models, namely, long-linear, quasi-logistic, and Gompertz curves, in modeling car ownership in Egypt. The results indicate that car ownership is influenced by time and GDP per capita. Comparing the results of the three models, he found that the Gompertz model has the highest forecasting accuracy than the other two models. It shows that by year 2024 car ownership will reach 33, 39, 47, 56, and 68 per 1,000 persons under 0%, 1%, 2 %, 3%, and 4 % income growth scenarios, respectively.

Li et al. (2010) applied both aggregate (linear regression) and disaggregate (multinational logit choice model) models to examine private car ownership across megacities in China. The results suggest that private car ownership is determined by urban affluence, urban scale, road infrastructure supply, and population density. They called for caution in the application of the results because there exist other variables other than the ones employed in the study that can influence car ownership.

Shaygan et al. (2017) reviewed existing studies on car ownership models and determinants in Iran. They found that the accuracy of the existing regression and discrete choice models is hindered by reliable aggregate and disaggregate data. They suggest that both disaggregate and aggregate data should be used to develop model types that will take cognizance of behavioral and economic variables. De Jong et al. (2004) used 16 criteria to compare the nine existing car ownership model types found in the literature. They found that the “most preferred model type will vary from context to context.”

Studies on the determinants of vehicle ownership in Africa are rare in literature. Most of the existing studies focused on Europe, America, South-east Asia, and Australia. This is even though Nigeria and most African countries are currently experiencing phenomenal growth in vehicle ownership. This article provides evidence from Nigeria for modeling car ownership in African countries

Theory

Existing models for determining vehicle ownership were categorized by Potoglou and Susilo (2008) into two broad groups, namely, aggregate and disaggregate models.

Aggregate Models

Aggregate models focus on the accumulation of household vehicle ownership at different geographical scales such as zonal, regional, state, or national level (Potoglou & Susilo, 2008). They can be applied to derive income elasticity of vehicle demand, make international comparisons, and predict future vehicle stock and can be used as input in forecasting travel demand applying the least squares method of regression techniques (Zegras & Hannan, 2012).

Time series, cohort survival, and car market analysis belong to this group of models (Dargay & Gately, 1999; European Commission [DGII] et al. 1999; Madre & Pirotte, 1991). The time series aggregate model uses an S-shaped curve and saturation level concept to develop vehicle ownership over time. They assume that vehicle ownership is influenced by income and GDP. Initially, the influence starts slowly and later rises rapidly to the optimum (De Jong et al., 2004). Some applications of aggregate time series models are the works of Whelan (2001), Whelan et al. (2000), Ingram and Liu (1998), and Dargay and Gately (1999). They used the models to explain vehicle ownership in some developed countries. The application of aggregate time series model is appealing to developing countries due to its small amount of data requirement (De Jong et al., 2004).

The aggregate cohort model divides the current population into 5-year age cohorts. The cohorts are then projected into the future, explaining their vehicle ownership concerning how they will acquire, keep, and lose cars as they become older (De Jong et al., 2004). It is most suitable for forecasting the effect of demographic dynamics on vehicle ownership. The most notable application of the model is Madre and Pirotte (1991) and Van den Broecke/Social Research (1987). They used it to examine cohort effect as a major cause of significant growth in car ownership in France and the Netherlands, respectively.

The aggregate car market model operates by determining the equilibrium between the forces of supply and demand. The demand side is the number of cars purchased as predictable by socioeconomic factors. On the contrary, the supply side is defined by the number of scrapped cars, aging, and new cars bought in the previous year. The rate of cars bought (supply function) and the cost of used cars (demand function) are the dependent factors in the equation, (Cramer & Vos, 1985). The change in the car market is influenced by the dynamics of demand for the available quantity of used cars. This operates through the cost of used vehicles as well as its effects on demand for new vehicles. Consequently, a unit decrease in the price of used cars results in a proportionate reduction in the request for new cars (De Jong, 2002). The examples of the application of the aggregate car market model include Berry et al. (1995), Cramer and Vos (1985), Manski (1983), and Mogridge (1983).

Disaggregate Models

Disaggregate models examine vehicle ownership at household levels (Potoglou & Susilo, 2008). The advantage of disaggregate models over aggregate models lies in their behavioral structure and enhanced ability in identifying a causal relationship. This has made disaggregate models the most dominant in determining vehicle ownership. Their extensive use was based on their robustness in developing the vehicle ownership model. This is because the models have overcome the weakness of aggregate models (Potoglou & Susilo, 2008). Disaggregate models use the ordinal or nominal discrete variables to determine vehicle ownership, thus giving rise to two types of model choice, namely, ordered and unordered. The ordered choice model is based on an assumption that the household’s choice of the number of vehicles is dependent on the propensity of a household to own vehicles. The unordered response model, on the contrary, assumes that the utility value of household vehicle ownership at different levels is determined by the one that gives the maximum utility. No study yet has proved that ordered response or unordered response models are the most appropriate (Bhat & Pulugurta, 1998; Potoglou & Kanoroglou, 2007; Potoglou & Susilo, 2008). Examples of disaggregate models include pseudo-dynamics, joint discrete, static disaggregate car-type choice, and static disaggregate car ownership (De Jong et al., 2004).

The relevance of these theories to this study cannot be overemphasized. They were very critical to the choice of model specification applied in this study. The aggregate time series model became the most appropriate model for this study due to the dearth of disaggregate data in the country. This challenge has made the use of disaggregate models impossible in this work. Despite the challenges, it is hoped that this study will be the foundation for future research. This is basically because its prediction of future vehicle ownership would serve as a foundation for future studies which will require a broader and more complete data set (Cirillo & Xu, 2011). This will be particularly useful for developing countries where studies on vehicle ownership at both aggregate and disaggregate levels are limited.

Method

Data used in the study which consists of socioeconomic and physical factors that influence vehicle ownership were collected from secondary sources. The variables are aggregated time series data at the national level covering a period of 40 years, from 1970 to 2010. These include stock of vehicle, per capita income, GDP, inflation rate, the price of fuel, accident cases, literacy level, and unemployment. Others are population, the rate of urbanization, length of roads, and stock of public transport vehicles. They were mainly collected from the Central Bank of Nigeria and the National Bureau of Statistics.

Due to the absence of disaggregated institutional household data in the country (the available household data were aggregated at the national level), this study was compelled to use multiple linear regression technique (an aggregate time series model). It was used to determine the socioeconomic and physical factors that influence vehicle ownership. The outcome of this model will be the basis for other more detailed models in future studies in Nigeria. The dependent variable (Y) was the vehicle ownership, whereas the independent variables (Xs) were represented by socioeconomic and physical factors.

The formula for multiple linear regression (stepwise method) as used in the study is given as

\begin{array}{l} V O = a + b_{1} g d p_{1} + b_{2} p c i_{2} + b_{3} i r_{3} + b_{4} f p_{4} + b_{5} a c_{5} + b_{6} l_{6} \\ + b_{7} u_{7} + b_{8} u r_{8} + b_{9} p_{9} + b_{10} l r_{10} + b_{11} s p t_{11} + e \end{array}

(1)

where VO is the vehicle ownership (dependent variable), gdp is the gross domestic product (independent variable), pci is the per capita income, ir is the inflation rate, fp is the fuel price, ac is the accident cases, l is the literacy level (number of people with the First School Leaving Certificate), u is the unemployment rate, ur is the urbanization rate, p is the population, lr is the length of road, spt is the stock of public transport vehicles, and a, b, and e are the constant of the model, regression coefficient, and standard error, respectively.

Data Presentation and Analysis

Table 1 shows the time series data on vehicle ownership and its determinant factors from 1970 to 2010 (40 years). Vehicle ownership grew from 32,537 in 1970 to 124,301 in 2010, thus representing an increase of 382%. The GDP, per capita income, unemployment, stock of public vehicles, and literacy rate increased, whereas urbanization was stable. However, accident cases and inflation rate had a ripple growth; they fluctuated from 1970 to 2010. The nation has experienced changes in fuel price 17 times since 1970. The population grew by 254% from 1970 to 2010, depicting a geometrical progression. There was a gradual increase in the length of roads constructed from 1970 (27,000 km) to 2010 (245,094 km) which represents a 908% increase.

Table 1.

Vehicle Ownership and the 11 Determinant Factors.

Year	Vehicle ownership (stock of vehicles)	Gross domestic product (billion naira)	Per capita income	Inflation rate	Fuel price	Accident cases	Literacy	Unemployment rate	Urbanization rate	Population	Length of road	Stock of public transport
1970	32,537	42.1	242,258.70	13.8	0.0915	23,320	5,193,550	3.1	3.750	62,642,390	27,000	12,779.00
1971	53,863	47.1	270,382.90	16.0	0.0915	17,745	5,950,297	4.9	3.751	63,707,310	27,000	21,017.00
1972	62,276	48.9	272,944.30	3.2	0.0915	23,287	8,242,060	2	3.750	64,790,335	27,000	23,615.00
1973	68,685	53.1	280,715.30	5.4	0.0915	24,844	9,834,827	3.2	3.750	65,891,770	27,000	24,669.00
1974	75,972	159.1	304,117.10	13.4	0.0915	28,894	11,410,360	6.2	3.751	67,011,930	27,000	30,232.00
1975	114,234	271.7	280,715.30	33.9	0.0915	32,651	12,749,403	4.1	3.750	68,151,133	28,000	43,117.00
1976	198,849	291.4	297,132.00	21.2	0.0915	53,897	19,589,875	4.3	3.750	69,309,703	28,000	119,605.00
1977	273,855	315.2	305,694.20	15.4	0.0915	35,841	13,760,030	2.1	3.750	70,487,968	28,000	147,533.00
1978	236,866	292.1	279,425.40	16.6	0.0915	36,121	14,285,437	1.6	3.740	71,686,263	28,000	133,680.00
1979	234,526	299.4	289,536.80	11.8	0.1315	NA	14,574,523	2	3.782	72,904,929	28,000	120,000.00
1980	295,472	315.4	293,200.60	9.9	0.1315	32,138	14,383,487	1.9	3.716	74,144,313	28,000	134,236.00
1981	386,748	94.3	247,876.90	20.9	0.1315	35,906	13,612,765	4.1	3.759	75,404,767	28,000	228,443.00
1982	224,150	101.0	238,954.80	7.7	0.1315	37,094	12,914,870	4.2	3.754	76,686,648	28,000	39,855.00
1983	140,411	110.1	221,196.50	23.2	0.1315	33,029	13,260,105	5.3	3.762	77,990,321	28,000	23,430.00
1984	112,972	116.3	211,302.80	39.6	0.1315	28,888	13,313,577	7.9	3.728	79,316,156	28,000	10,078.00
1985	104,864	134.6	223,088.30	5.5	0.1315	28,976	13,740,505	6.1	3.760	80,664,531	28,000	9,629.00
1986	58,595	134.6	198,319.60	5.4	0.345	25,188	13,607,249	5.3	3.749	82,035,828	28,000	5,790.00
1987	40,145	193.1	172,402.70	10.2	0.345	24,206	13,776,854	7.0	3.748	83,430,437	32,000	4,523.00
1988	21,097	263.3	180,584.50	38.3	0.345	25,792	14,805,937	5.1	3.756	84,848,754	32,000	3,400.00
1989	7,934	382.3	187,298.50	40.9	0.43	23,242	15,870,280	4.5	3.748	86,291,183	32,000	1493.00
1990	9,303	382.6	205,824.70	7.5	0.575	21,827	16,190,947	3.5	3.748	87,758,133	32,000	1547.00
1991	37,220	545.7	199,405.90	13.0	0.575	65,658	15,741,078	3.1	3.757	88,992,220	32,000	7,416.00
1992	48,757	875.3	195,279.50	44.5	0.575	22,489	14,078,273	3.4	3.732	91,510,699	32,179.86	7,243.00
1993	78,716	1089.7	194,427.80	57.2	3.125	21,734	17,397,553	2.7	3.768	94,100,452	NA	4,565.00
1994	62,119	1399.7	191,358.20	57.2	10	18,523	16,945,567	2.0	3.742	96,763,495	31,097.05	11,040.00
1995	67,918	2907.4	186,069.00	72.8	10	15,828	17,907,008	1.8	3.746	99,501,902	32,097.05	18,958.00
1996	87,027	4,032.3	190,545.70	29.3	10	14,718	19,158,541	3.4	3.761	102,339,270	32,079.00	21,903.00
1997	124,398	4,189.2	191,055.20	8.5	10	17,537	19,263,534	3.2	3.747	105,339,270	32,098.00	56,430.00
1998	141,710	3,989.5	186,069.00	10.0	17	17,179	19,861,654	3.2	3.745	108,550,651	34,123.00	57,524.00
1999	264,432	4,679.2	190,545.70	6.6	21.5	15,864	25,765,969	3.1	3.751	111,796,316	34,123.95	72,049.00
2000	216,173	6,713.6	191,055.20	6.9	21.5	16,990	21,575,178	3.7	3.743	126,638,296	NA	77,648.00
2001	144,380	6,895.2	191,397.70	18.9	25	20,520	22,099,553	3.8	3.764	130,412,118	NA	51,819.00
2002	173,400	7,795.8	187,546.10	12.9	25	14,544	21,717,789	12.60	3.747	134,298,399	NA	62,252.00
2003	207,400	9,913.5	192,616.40	14.0	33	14,363	20,469,395	14.80	3.748	138,300,491	34,340.95	74,456.00
2004	217,600	11,411.1	198,437.80	15.0	56.5	14,279	21,294,518	13.35	3.753	142,421,846	34,340.95	78,119.00
2005	227,580	14,610.9	213,475.70	17.9	58	18,659	18,980,395	11.9	3.743	146,666,017	34,341.25	81,689.00
2006	357,104	18,564.6	278,249.00	8.5	65	18,662	20,663,805	12.3	3.765	140,431,790	34,341.25	175,284.00
2007	510,267	20,657.3	280,457.70	5.4	70	17,797	5,193,550	12.7	3.733	144,897,327	36,183	260,356.00
2008	85,297	24,296.3	295,636.10	15.1	65	17,854	5,950,297	14.9	3.763	149,510,846	36,183	60,900.00
2009	80,246	24,712.7	307,593.60	13.9	65	13,703	8,242,060	19.7	3.735	154,277,742	36,183	57,393.00
2010	124,301	33,984.8	318,307.70	11.8	65	20,910	9,834,827	21.4	3.763	159,203,664	36,183	95,597.00

Source. Central Bank of Nigeria and National Bureau of Statistics.

Note. NA = not available.

The results in Table 2 suggest that the socioeconomic and physical factors have a significant effect on vehicle ownership in the country at .01 level and adjusted explained variance of 0.931.

Table 2.

The Parameters: Linear Regression Results.

R = .965	F-cal = 94.797
R² = .931	α–significant = .000
Adjusted R² = .921	p < .01
Standard error = 48,281

Table 3 shows that five of the 11 identified factors, namely, GDP, stock of public transport vehicles, fuel price, per capita income, and literacy level, are the determinants of vehicle ownership. However, the remaining six variables were not significant. These are population, inflation rate, accident cases, unemployment, urbanization, and length of the road. Furthermore, the collinearity test shows that the variance inflation factor (VIF) was less than 10, and this indicates the absence of high multicollinearity between the predictor variables (i.e., the independent variables were not correlated). This suggests that the model of this study is reliable.

Table 3.

The Relationship Between Vehicle Ownership and Each of the 11 Socioeconomic and Physical Factors.

			Collinearity statistics
Variable	Standardized coefficient (β)	T	Tolerance	VIF	Significance level	Comment
Constant		−3.584			.001	Significant
Gross domestic product	1.023	6.0305	.876	1.142	.000	Significant
Stock of public transport vehicles	0.300	4.675	.478	2.033	.000	Significant
Fuel price	−0.607	−3.703	.101	9.934	.001	Significant
Per capita income	0.339	4.238	.307	3.258	.000	Significant
Literacy	0.286	3.221	.250	4.001	.003	Significant
Population	−0.151	−0.574	.221	4.526	.570	Not significant
Inflation rate	−0.001	−0.025	.831	1.203	.980	Not significant
Accident cases	0.008	0.144	.590	1.696	.886	Not significant
Unemployment rate	0.099	1.181	.277	3.616	.246	Not significant
Urbanization rate	−0.002	−0.050	.958	1.043	.960	Not significant
Total length of roads	−0.208	−1.366	.322	3.106	.181	Not significant

Note. VIF = variance inflation factor.

Discussion

The analysis of result of the hypothesis suggests that there was a very strong relationship between vehicle ownership and four socioeconomic and one physical factor initially identified in this research (see Table 2). These are GDP, literacy level, per capita income, fuel price, and stock of public transport vehicles (see Table 3). The aforementioned five factors accounted for 93.1% growth in vehicle ownership. However, the other remaining six variables, namely, population, inflation, length of roads, accidents, unemployment, and urbanization, contributed insignificantly to the growth in vehicle ownership (see Table 3).

The predictive model is therefore as follows:

\begin{array}{l} V O = - 3.58 + 1.023 (g d p) + 0.300 (s p t) + - 0.607 (f p) \\ + 0.339 (p c i) + 0.286 (l) + 48, 281 \end{array}

(2)

where VO is the vehicle ownership, a is the constant of the model, gdp is the gross domestic product (socioeconomic factor), spt is the stock of public transport vehicles (physical factor), fp is the fuel price (socioeconomic factor), pci is the per capita income (socioeconomic factor), and l is the literacy rate (socioeconomic factor).

Consequent upon the application of the predictive model, vehicle ownership was 35.3 million vehicles in 2018. It will increase to 48.7, 66.2, and 76.1 million in 2030, 2040, and 2050, respectively. This shows that the vehicle population will rise by 215.7% from 2018 to 2050. The rapid growth in vehicle ownership poses a serious challenge for energy security, road infrastructure, environment, traffic movement, road safety, and economy. Vehicle ownership per person was 0.174 in 2018. It will rise slightly to 0.176 and 0.185 in 2030 and 2040, respectively. However, a decrease to 0.164 will be experienced in 2050. The vehicle population per 1,000 persons was 174 in 2018. This will increase to 176.1 in 2030 and 185.3 in 2040, whereas a decrease to 163.7 will be recorded in 2050. This pattern of growth is similar to China’s vehicle growth pattern which depicts the S-shaped curve (Wu et al., 2014). However, vehicle ownership in Nigeria will constitute 12.2% of China, 19% of the United States, and 2.5% of the world vehicle population in 2050. It was 29 times larger than the vehicle population in the developing countries of Tanzania, Guinea, Sierra Leone, Chad, Rwanda, and Burundi, which were shown to have vehicle ownership of six vehicles per 1,000 persons in 2018.

The implication of the results with regard to each of the respective variables is as follows: The GDP was significantly related to growth in vehicle ownership, (β = 1.023; t = 6.0305; p = .000 [<.01 significance level]). The elasticity suggests that a 1% increase in GDP will result in 1.023% increases in vehicle ownership. This reveals that vehicle ownership responds directly to the dynamics of changes in GDP. The two factors move in the same direction, implying that the higher the GDP, the higher the stock of vehicles, ceteris paribus. This is an indication that GDP is a major factor that contributes to the phenomenal increase in vehicle ownership in the country. The outcome of this result is consistent with Wu et al. (2014) and Dargay et al. (2007) which showed that GDP has a significant effect on growth in vehicle ownership in China and some selected countries.

The influence of the stock of public transport vehicles on the dynamics of vehicle ownership in the country was significant (β = 0.300; t = 4.675; p = .000 [<.01 significance level]). The elasticity indicates that a 1% increase in the stock of public transport vehicles will result in a 0.300% increase in vehicle ownership. This implied that vehicle ownership increases with the growth in the stock of public transport vehicles. This differs from the works of Eriksson et al. (2010), Liu et al. (2011), and Nedal et al. (2018) which found that improved public transportation system results in significantly less private vehicle ownership and use.

The relationship of the price of fuel to the growth of vehicle ownership was negative and significant (β = −0.607; t = −3.703; p = .001 [<.01 significance level]). The elasticity indicates that a 1% decrease in fuel price results in a 0.607% increase in vehicle ownership. The implication is that people are motivated to own vehicles when fuel prices are low, thus resulting in an increase in vehicle ownership. The finding agreed with the work of Tsang et al. (2011) which showed that fuel price has a significant effect on vehicle ownership in Sydney, Australia.

Per capita income was significantly related to vehicle ownership in the country (β = 0.339; t = 4.238; p = .000 [<.01 significance level]). Vehicle ownership responds directly to the dynamics of per capita income. The elasticity of demand showed that a 1% increase in per capita income will cause 0.339% increases in vehicle ownership. The two factors moved in a similar direction, thus conforming to the law of supply and demand which states that an increase in income leads to an increase in demand. This indicates that when income increases, people will tend to be motivated to own or change vehicles, thereby increasing the stock of vehicles. The indication is that per capita income is a major factor that contributes to the phenomenal increase in vehicle ownership in the country. This result was consistent with the work of Tsang et al. (2011), Dargay et al. (2007) and De Jong et al. (2004), which found that income had a significant effect on vehicle ownership in their respective studies.

The literacy level was also significantly related to vehicle ownership (β = 0.286; t = 3,221; p = .003 [<.01 significance level]). The elasticity showed that a 1% increase in literacy level will result in a 0.286% increase in vehicle ownership. This implies that vehicle ownership responds positively to the dynamics of literacy. This suggests that more literate people owe more vehicles than less literate people.

Although the population is one of the variables used in existing studies in some countries, the results show that it was not significantly related to the growth in vehicle ownership in Nigeria (β = −0.151; t = −.574; p = .570 [>.05 significance level]). This indicates that demand for vehicles does not significantly change with an increase in population. This is quite contrary to the situation in more buoyant economies. The reason is that about 70% of Nigerians live below the poverty level (less than US$1 per day). The outcome of this result agrees with that of Dargay and Gately (1999) which showed that the vehicle population is lower in densely populated urban areas occupied by low-income groups.

The inflation rate is not significantly related to growth in vehicle ownership (β = −0.001; t = −0.025; p = .980 [>.05 significance level]). This suggests that the increase or decrease in the inflation rate does not significantly influence changes in the demand for vehicles in the country.

Vehicle ownership was not significantly responsive to the changes in road traffic accidents (β = 0.141; t = 0.008; p = .886 [>.05 significance level]). This indicates that the number of accidents on the roadways does not have a significant effect on growth in vehicle ownership. It could be attributed to the fact that the mobility of some people in Nigeria does not depend on vehicles only. They have such other alternatives as tricycles, motorcycles, bicycles, and walking. This finding differed from that of Ukoji (2014) which found that road traffic accident causes a decrease in vehicle ownership in the country.

The unemployment rate was not significantly related to the growth in vehicle ownership (β = 1.181; t = 0.099; p = .256 [>.05 significance level]). This suggests that the unemployment rate does not have a significant influence on vehicle ownership. This is surprising because, despite the high unemployment rate, vehicle ownership has continued to experience a phenomenal rise in the country. This is probably because vehicle ownership increase was propelled by fuel price subsidy that distorted the market structure in favor of the employed people who had money to spend.

The urbanization rate was not significant to vehicle ownership in the country (β = −0.050; t = −0.002; p = .960 [>.05 significance level]). This suggests that the increase or decrease in urbanization rate does not significantly influence growth in vehicle ownership. This outcome was confirmed by Dargay et al. (2007) who showed that urbanization does not have a significant effect on vehicle ownership worldwide. He opined that countries that experienced a higher rate of urbanization did not seem to have a greater number of vehicle ownership than others.

The length of roads was also not significantly related to the growth in vehicle ownership (β = −1.366; t = −0.208; p = .181 [>.05 significance level]). This suggests that the length of roads constructed does not significantly influence the demand for vehicles in the country. The outcome of this result does not agree with the work of Dargay and Gately (1999) which found that for a given level of car ownership, there was a wide range of road densities, especially for higher income countries. However, it concurred with the work of Ingram and Liu (1998) which argued, through the examination of data from selected countries, that investment in roads was strongly associated with economic growth and not vehicle ownership.

Overall, the GDP had the most significant effect on vehicle ownership. It was followed in descending order by fuel price, per capita income, stock of public transport vehicles, and literacy level (see Figure 1). A comparative analysis of the result of this study with existing studies found in Europe, America, South-east Asia, and Australia confirmed that GDP, fuel price, per capita income, and stock of public transport are factors that drive vehicle ownership around the globe. In the same vein, existing studies found in the Middle East and North African (MENA) countries of Iran, Turkey, Kuwait, and Egypt confirmed income, fuel price, and GDP as determinants of vehicle ownership globally (Ceylan et al., 2018; Kermanshah & Ghazi, 2001; Said, 1992; Soltani, 2017). However, literacy level as a factor of vehicle ownership was peculiar to Nigeria alone, whereas population, urbanization, length of the road, unemployment, and inflation were peculiar to the comparative continents. This is except for MENA which has the household size, number of employees, occupation, urban form, and age of household’s head as peculiar factors that determine vehicle ownership. Other studies conducted in sub-Saharan African countries of Burkina Faso, Rwanda, Cameroon, Gabon, and Mauritius found GDP accountable for vehicle ownership (Button et al., 1992). This indicates that there are particular factors (apart from the global factors) that drive vehicle ownership in Nigeria. Furthermore, apart from GDP and per capita income (which are already known), three new variables influence vehicle ownership in the country. These are the stock of public transport vehicles, price of fuel, and literacy level.

Figure 1.

Determinants of vehicle ownership in Nigeria.

Recommendations

The five factors, namely, GDP, per capita income, fuel price, literacy level, and stock of public vehicles, are confirmed determinants of vehicle ownership in Nigeria. Three new variables (the stock of public transport vehicles, price of fuel, and literacy level), among the five factors, were found to have influenced vehicle ownership. The vehicle population was 35.3 million units in 2018. This will increase to 48.7, 66.2, and 76.1 million units in 2030, 2040, and 2050, respectively. The aforementioned new facts should be the bases for the formulation of appropriate policy on vehicle ownership.

The policy should also address the likely challenges posed by the growth in vehicle ownership on energy security, road infrastructure, environment, traffic movement, road safety, and economy. Therefore, the policy should consider the following cardinal objectives to effectively tackle such challenges. First is the need to expand the existing road infrastructure to accommodate more vehicles. This will reduce traffic congestion and road traffic accidents that are associated with growth in vehicle ownership. Second is the development of alternative modes of transportation that have the capacity of carrying a large number of commuters such as metro lines, monorail, tram, and ferry services to supplement the use of private vehicles. Third is to ensure the development of measures that will improve the quality of public transportation. This is to enhance the safety and comfort of passengers as well as discourage automobile dependency. Finally, encourage the promotion of energy transition to clean and renewable energy to reduce the dependence on fossil fuel and minimize the effect of climate change arising from automobile carbon dioxide emissions on the environment. This will save the nation’s economy from the huge burden of fuel subsidy.

Conclusion

The results of this study suggest that four socioeconomic factors (GDP, per capita income, fuel price, literacy level) and one physical factor (stock of public transport vehicles) were the determinants of vehicle ownership. Three new variables, among the five factors, that were found to influence vehicle ownership are the stock of public transport vehicles, price of fuel, and literacy level. The GDP was the most important factor that influenced vehicle ownership. It was followed in descending order by fuel price, per capita income, stock of public transport vehicles, and literacy level. The stock of vehicles was 35.3 million in 2018. It will increase to 48.7, 66.2, and 76.1 million in 2030, 2040, and 2050, respectively. Consequently, the need to formulate appropriate policy on vehicle ownership based on the new facts is inevitable.

Footnotes

Data Accessibility Statement

The data used in this study are already deposited in 4TU.Centre for Research Data. The link is

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Maxwell Umunna Nwachukwu

References

Agbonkhese

Yisa

G. L.

Agbonkhese

E. G.

Akanbi

D. O.

Aka

E. O.

Mondigha

E. B.

(2013). Road traffic accidents in Nigeria: Causes and preventive measures. Civil and Environmental Research, 3(13), 90–99.

Berri

(2009). A cross-country comparison of household car ownership: A cohort analysis. IATSS Research, 33(2), 21–38.

Berry

Levinsohn

Pakes

(1995). Automobile prices in market equilibrium. Econometrica, 63(4), 841–890.

Bhat

Pulugurta

(1998). A comparison of two alternative behavioral choice mechanisms for household auto ownership decisions. Transportation Research Part B: Methodological, 32(1), 61–75.

Button

Hine

Ngoe

(1992). Car ownership forecasts for low-income countries. Traffic Engineering and Control, 33(12), 666–671.

Cao

Huang

(2013). City-level determinants of private car ownership in China. Asian Geographer, 30(1), 37–53.

Ceylan

Basken

Ozan

(2018). Modeling and forecasting car ownership based on socio-economic and demographic indicators in Turkey. TeMA—Journal of Land Use, Mobility and Environment, Special issue 1, 47–66. https://doi.org/10.6092/1970-9870/5452

Cirillo

(2011). Dynamic discrete choice models for transportation. Transport Reviews, 31(4), 473–494.

Cramer

J. S.

Vos

(1985, May). A model for passenger car forecasts. Interfaculty of Actuarial Science and Econometrics, University of Amsterdam.

10.

Dargay

(2002). Determinants of car ownership in rural and urban areas: A pseudo-panel analysis. Transportation Research Part E: Logistics and Transportation Review, 38, 351–368.

11.

Dargay

(2007). The effects of prices and income on car travel in the UK. Transportation Research Part A: Policy and Practice, 41, 949–9604.

12.

Dargay

Gately

(1999). Income’s effect on car and vehicle ownership worldwide. Transportation Research Part A: Policy and Practice, 33(2), 101–138.

13.

Dargay

Gately

Sommer

(2007). Vehicle ownership and income growth, worldwide 1960–2030. The Energy Journal, 28(4), 143–170.

14.

Das

Dutta

Sharifuddin

(2010). Car ownership growth in Delhi. Decision, 37(2), 51–61.

15.

De Jong

Fox

Pieters

Daly

A. J.

Smith

. (2004). A comparison of car ownership models. Transport Reviews, 24(4), 397–408.

16.

Ding

Wang

Yang

Liu

Lin

(2016). Spatial heterogeneous impact of built environment on household auto ownership levels: Evidence from analysis at traffic zones scales. Transportation Letters, 8(1), 26–34.

17.

Eakins

(2015). Household car ownership in Ireland across time: An analysis of the effects of expansion and contraction in the Irish economy. International Journal of Transport Economics, 42(3), 315–342.

18.

Eriksson

Firman

Etkema

Fujji

Garlong

(2010). Experimental simulation of car users switching to public transportation. Transportation Letters, 2(3), 145–155.

19.

European Commission (DGII), Standard & Poors DRI, & CES-KU Leuven. (1999). Auto-oil II cost-effectiveness, Part II: The TREMOVE Model 3. http://ec.europa.eu/environment/enveco/air/index.htm

20.

Gomez-Gelvez

J. A.

Obando

(2013). Modeling car ownership in urban areas of developing countries: Case study of Bogoota, Colombia. Transportation Research Record, 2394(1), 111–118.

21.

Hao

Wang

(2011). Hybrid modeling of China’s vehicle ownership and projection through 2020. Energy, 36(2), 1351–1361.

22.

Hirota

(2007). Passenger car ownership estimation toward 2030 in Japan—BAU scenario with socio-economic factors. Studies in Regional Science, 37(1), 25–39.

23.

Huang

(2011). Car ownership modeling and forecasts for China [Open access paper]. Michigan Technological University.

24.

Ingram

Liu

(1998). Determinants of motorization and road provision. Research Advisory Staff and the Transport, Water, and Urban Development Department.

25.

Joseph

O. O.

Eromietse

E. J.

Emmanuel

Olufunke

(2017). Multinomial logit model estimation of household characteristics influencing car ownership in Akure, South West Nigeria. International Journal for Traffic and Transport Engineering, 7(2), 203–215.

26.

Kermanshah

Ghazi

(2001). Modeling automobile ownership decisions. A disaggregate approach. Scientia Iranica, 8(1), 29–37.

27.

Kobos

P. H.

Erickson

J. D.

Drennen

T. E.

(2003). Scenario analysis of Chinese passenger vehicle growth. Contemporary Economic Policy, 21(2), 200–217.

28.

Korkmaz

Akgungor

A. P.

(2018). The estimation of vehicle ownership via flower pollination algorithm. Gazi Muhendislik Bilunleri Dergisi, 4(1), 39–45.

29.

Le Vine

Polak

. (2017). A nationwide study of factors associated with household car ownership in China. IATSS Research, 42(3), 128–137. https://doi.org/10.1016/j.iatssr.2017.10.001

30.

Walker

J. L.

Srinivasan

Anderson

W. P.

(2010). Modeling private car ownership in China: Investigation of urban form impact across mega cities. Transportation Research Records, 2193, 76–84. https://doi.org/10.3141/2193-10

31.

Liu

Tremblay

J. M.

Cirillo

(2011). A model system for household vehicle holding, type and usage based on 2009 NHTS. University of Maryland, Department of Civil and Environmental Engineering.

32.

Sun

Wang

(2017). Analysis and prediction on vehicle ownership based on improved stochastic Gompert Duticusco process. Journal of Advanced Transportation, 27, 1–6. https://doi.org/10.1155/2017/4013875

33.

Madre

Pirotte

(1991). Regionalization of projections fleet and car traffic. Methods of analysis of displacement behaviors for the 1990s. In 6th international conference on travel behavior (pp. 413–430). Quebec Association of Transport and des Routes, Château Bonne Entente, Quebec.

34.

Manski

C. F.

(1983). Analysis of equilibrium automobile holdings in Israel with aggregate discrete choice models. Transportation Research B: Methodological, 17(5), 373–389.

35.

Martha

Y. M.

Kroesen

Vanwee

Van Daalen

(2017). Changing influence of factors explaining household car ownership levels in the Netherland Transportation Research Records, 2666, 103–111. https://doi.org/10.3141/2666-12

36.

Matas

Raymond

(2008). Changes in the structure of car ownership in Spain. Transportation Research Part A:: Policy and Practice, 42, 187–202.

37.

McCarthy

Wang

(2017). Private vehicle ownership in provisional China. Journal of the Transportation Research Forum, 58(3), 91–109.

38.

Mogridge

M. J. H.

(1983). The car market. Pion.

39.

National Bureau of Statistics. (2018). Annual abstract of statistics 1970–2017.

40.

Nedal

T. R.

Gazdar

Assi

R. J.

(2018). Effect of public transportation in reducing passenger car trips to schools in Al-Khobar–Dhahran Metropolitan Area, Saudi Arabia. Transportation Letters, 10(1), 43–51.

41.

Nwachukwu

M. U.

Chike

(2011). Fuel subsidy in Nigeria: Fact or fallacy. Energy, 36(5), 2796–2801.

42.

Ogut

K. S.

(2004). S-curve models to determine the car ownership in Turkey. ARI, The Bulletin of the Istanbul Technical University, 54(2), 1–4.

43.

Pasra

Ramli

M. I.

Dharmowijoyo

D. E.

(2018). Car ownership of households in sub-urban area in Makassar city. In Raza

Mustafa

(Eds.), MATEC Web of Conferences 203,05005, ICCOEE (pp. 1–9). EDP Sciences.

44.

Potoglou

Kanaroglou

P. S.

(2007). Household demand and willingness to pay for clean vehicles. Transportation Research Part D: Transport & Environment, 12(4), 264–274.

45.

Potoglou

Susilo

Y. O.

(2008). Comparison of vehicle-ownership models. Transportation Research Record, 2076, 97–105.

46.

Prabnask

Taylor

M. A. P.

Yue

W. L.

(2011). Modelling household vehicle ownership in Asian medium-sized urban area: A case study of Khon Kaen City. Thailand. Journal of the Eastern Asia Society for Transportation Studies, 9, 196–210.

47.

Ritter

Vance

(2012). Do fewer people mean fewer cars? Population decline and car ownership in Germany. Transportation Research Part A: Policy and Practice, 50, 74–85.

48.

Romilly

Song

(2000). Modelling and forecasting car ownership in Britain: A cointegration and general to specific approach. Journal of Transport Economics and Policy, 32(2), 165–185.

49.

Said

G. M.

(1992). Modeling household car ownership in the Gulf States. Journal of Transport Economics and Policy, 26(2), 121–138.

50.

Serag

M. S.

(2014). Modeling car ownership in Egypt. The Egyptian International Journal of Engineering Sciences and Technology, 17(2), 1812–1819.

51.

Shaygan

Mamdoohi

Masoum

H. E.

(2017). Car ownership models in Iran: A review of methods and determinants. Transport and Telecommunication, 18(1), 45–59.

52.

Soltani

(2017). Social and urban form determinants of vehicle ownership: Evidence from a developing country. Transportation Research Part A: Policy and Practice, 96, 90–100.

53.

Song

Wang

(2017). Measuring the spatial dimension of automobile ownership and its associations with household characteristics and land use patterns. A case study in three countries, South Florida, USA. Sustainability, 9(4), Article 558.

54.

Tam

Lam

W. H. K.

(1999). Modelling car ownership in Hong Kong: A discrete choice approach. Journal of Asia Society for Transportation Studies, 3(5), 319–332.

55.

Tsang

Daley

Milthorpe

(2011, September 28–30). Forecasting car ownership in the Sydney area. In Young

Dolman

(Eds.), Australian Transport Research Forum proceedings (pp. 1–14). Adelaide, Australia: EDP Sciences.

56.

Ukoji

V. N.

(2014). Trends and patterns of fatal road accidents in Nigeria (2006–2014) (pp. 1–45). IFRA-Nigeria working papers series, No. 35.

57.

Van den Broecke/Social Research. (1987). The possible growth of passenger car ownership until 2010: Report for PbIVVS. BSR.

58.

Wang

Huo

Johnson

(2007). Projection of Chinese motor vehicle growth, oil demand and CO₂ emission through 2050. Transportation Record, 2038, 69–77. https://doi.org/10.3141/2038-09

59.

Wards. (2014). Automotive industry news, data and statistics. https://www.loc.gov/rr/business/BERA/issue2/issue2main.html

60.

Whelan

(2001). Methodological advances in modelling and forecasting car ownership in Great Britain [Paper presentation]. European Transport Conference 2001, PTRC, Cambridge.

61.

Whelan

Wardman

Daly

(2000). Is there a limit to car ownership growth? An exploration of household saturation levels using two novel approaches [Paper presentation]. European Transport Conference 2000, PTRC, Cambridge.

62.

Wong

K. L.

(2013). An analysis of car and motorcycle ownership in Mexico. International Journal of Sustainable Transportation, 7(3), 204–225.

63.

Zhao

(2014). Vehicle ownership analysis based on GDP per capita in China: 1963–2050. Sustainability, 6, 4877–4899.

64.

Yagi

Managi

(2016). Demographic determinants of car ownership in Japan. Transport Policy, 50, 37–53.

65.

Yin

Shao

Wang

(2019). Exploring the impact of built environment on car use—Does living near urban rail transits matter? Transportation Letters, 9, 524–534. https://doi.org/10.1080/19427867.2019.1611196

66.

Zegras

P. C.

Hannan

V. H.

(2012). Dynamics of automobile ownership under rapid growth. Transportation Research Record, 2323, 80–89.

Determinants of Vehicle Ownership in Nigeria

Abstract

Keywords

Introduction

Literature Review

Theory

Aggregate Models

Disaggregate Models

Method

Data Presentation and Analysis

Discussion

Recommendations

Conclusion

Footnotes

Data Accessibility Statement

Declaration of Conflicting Interests

Funding

ORCID iD

References