Journal of Transport Geography 46 (2015) 122–128

Contents lists available at ScienceDirect

Journal of Transport Geography

journal homepage: www.elsevier.com/locate/jtrangeo

The motorcycle to passenger car ownership ratio and economic growth:

A cross-country analysis
Teik Hua Law ⇑, Hussain Hamid, Chia Ning Goh
Road Safety Research Centre, Faculty of Engineering, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia

a r t i c l e i n f o a b s t r a c t

Article history: Cross-country statistics have revealed steady growth in the number of motorcycles in many less
Received 15 April 2014 advanced economic countries (LAEC) with emerging economies due to increased urbanisation and
Revised 2 June 2015 personal wealth. In contrast, an opposite trend is occurring in advanced economic countries (AEC), with
Accepted 3 June 2015
cars replacing motorcycles as income grows. Motor vehicle crashes and injuries are an inevitable conse-
Available online 11 June 2015
quence of a high motorcycle population. This study focused on understanding how economic growth
affects the motorcycle to passenger car (MPC) ownership ratio and what factors underlie this relation-
Keywords:
ship. The data used in this analysis contained a sample of 80 countries at various levels of economic
Motorcycle ownership
Car ownership
developmental growth over the 48-year period between 1963 and 2010. The results pointed to an
Economic growth inverted U-shaped relationship between the MPC ownership ratio and the per capita Gross Domestic
Fixed effects panel linear regression Product (GDP). Generally, the MPC ownership ratio increased with income at a lower level and decreased
with income at a higher level. The evidence indicated that urbanisation, the total road length per thou-
sand population, and a proxy for purchasing power with regard to vehicle purchases were the underlying
factors that contributed to this relationship.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction in vehicle ownership, particularly for geographic regions during

The motorcycle is a major means of transportation in many less
advanced economic countries (LAEC), such as Vietnam, Malaysia
and Cambodia. This is due mainly to the affordability of
motorcycles, with more people in LAEC able to purchase them. It
is also due to their high manoeuvrability on congested roads.
Cross-country statistics have revealed steady growth in the num-
ber of motorcycles in many less advanced economic countries
(LAEC) with emerging economies due to increased urbanisation
and personal wealth. In contrast, an opposite trend is occurring
in advanced economic countries (AEC), with cars replacing motor-
cycles as income grows (Lai et al., 2006; Yamamoto, 2009; Chiou
et al., 2009; Pongthanaisawan and Sorapipatana, 2010). Motor
vehicle crashes and injuries are an inevitable consequence of a high
motorcycle population (Abdul Manan and Várhelyi, 2012; NHTSA,
2008; Preusser et al., 1995; Radin Umar et al., 1995; Ranney
et al., 2010; Sharma, 2008).
Vehicle ownership is typically influenced by socioeconomic
factors (De Jong et al., 2004) and can be modelled by aggregated
or disaggregated models. The aggregate model predicts changes
⇑ Corresponding author.
E-mail addresses: lawteik@upm.edu.my (T.H. Law), hushamid@upm.edu.my
(H. Hamid), gcnleonard@hotmail.com (C.N. Goh).
certain periods (Tanner, 1978; Khan and Willumsen, 1986; Button
et al., 1993; Dargay and Gately, 1999; Sillaparcharn, 2007; Ingram
and Liu, 1999). The advantage of this technique is that the models
do not require extensive survey data. The alternative model uses
disaggregate data to model vehicle ownership at the level of users
(Whelan, 2007). This model relates an individual’s propensity to
own a vehicle to various owners’ attributes, such as demographic,
socioeconomic, household and geographical characteristics, vehicle
price, and the availability of other transport modes (Delbosc, 2013;
Chiou et al., 2009; Nolan, 2010). Yet, due to the need to collect
extensive and detailed socioeconomic survey data, this model
structure has rarely been adopted in developing countries.
Over the last three decades, many researchers have investigated
the relationship between car ownership and income growth. Linear
and logarithmic functions were commonly employed to describe
the long-term growth of car ownership in earlier studies (Khan
and Willumsen, 1986). However, one of the limitations of these
models was that they led to unreasonable vehicle ownership
growth at higher income levels. In response to this limitation, a sig-
moid curve function with the saturation level of car ownership was
adopted (Tanner, 1983; Button et al., 1993; Ingram and Liu, 1999;
Dargay and Gately, 1999; Whelan et al., 2000; Whelan, 2001;
Pongthanaisawan and Sorapipatana, 2010; Sillaparcharn, 2007).
With this model, car ownership increases slowly at lower income

http://dx.doi.org/10.1016/j.jtrangeo.2015.06.007
0966-6923/Ó 2015 Elsevier Ltd. All rights reserved.
 T.H. Law et al. / Journal of Transport Geography 46 (2015) 122–128
levels, then increases rapidly, and finally approaches saturation
level.
Previous studies of the relationship between population density
and car ownership revealed that a higher spatial or urban popula-
tion density was associated with lower car ownership (Yamamoto,
2009; Lam and Tam, 2002; Clark, 2007, 2009; Khan and
Willumsen, 1986; Hess and Ong, 2002; Riley, 2002). This is probably
due to greater congestion, increased parking constraints and costs,
and a more efficient public transport system in densely populated
areas (Dargay and Gately, 1999; De Jong and Van de Riet, 2008;
Schwanen et al., 2004). Population density was found to have a pos-
itive effect on motorcycle ownership (RAND, 2004). According to
Wong (2013), this was due to motorcycles offering a more efficient
and cheaper means of transport than cars in population-dense areas.
Several studies demonstrated that increases in road density were
associated with growth in car ownership (Sehatzadeh et al., 2011;
Cao and Huang, 2013; De Jong and Van De Riet, 2008). Cao and
Huang (2013) indicated that a 1% increase in road density raised
car ownership by 0.13%. Motoring cost was also identified as a sig-
nificant determinant of car ownership, with studies showing that
higher motoring costs reduced car ownership (Dargay and
Vythoulkas, 1999; Dargay and Hanly, 2007; De Jong and Van de
Riet, 2008). Romilly et al. (1998) estimated that a 1% increase in
motoring cost decreased car ownership by 0.3% in the short term
and 2.2% in the long term. Previous empirical evidence revealed that
running cost had a greater impact than purchase costs on the elastic-
ity of car ownership (Whelan et al., 2000; Dargay and Gately, 1999).
The literature on motorcycle ownership is rather scarce, with
most studies conducted in East Asian countries, such as
Taiwan, Malaysia, Indonesia, Thailand, Japan and Vietnam
(Pongthanaisawan and Sorapipatana, 2010; Sillaparcharn, 2007;
Tuan, 2011; Tuan and Shimizu, 2005; Sanko et al., 2009; Hsu,
2005; Nagai et al., 2003; Dao and Duc, 2005). Previous evidence
revealed an inverted U-shaped relationship between motorcycle
ownership and income growth (Pongthanaisawan and
Sorapipatana, 2010; Sillaparcharn, 2007; Tuan, 2011; Senbil et al.,
2007; Nishitateno and Burke, 2014). Pongthanaisawan and
Sorapipatana (2010) indicated that as a country developed, motor-
cycle ownership increased but that it fell once income level
exceeded the threshold level. The study attributed this finding to
motorcycle ownership growing with the increasing demand for
transport in the early stages of economic growth. Eventually, prob-
ably due to the prestige, convenience, safety and comfort, people
shifted from motorcycles to car ownership as their income grew.
Although previous empirical studies have addressed the rela-
tionship between motorcycle ownership and income growth, there
is little explanation provided for the mechanisms by which income
growth leads to the inverted U-shaped relationship. The present
study focused on understanding how economic growth affects
the motorcycle to passenger car (MPC) ownership ratio and what
factors underlie this relationship. In particular, several variables
that are correlated with economic growth of a country, such as
urbanisation, road density and a proxy for purchasing power with
regards to vehicle purchases were used to explain this relationship.
The findings of this study aim to enhance the understanding of the
determinants of passenger car and motorcycle ownership and the
mechanisms affecting the growth of passenger car and motorcycle
ownership. This would allow policy makers to formulate more
effective transport policies and strategies.
2. Data and variable description
The data used in this analysis contained a sample of 80
countries at various levels of economic developmental growth over
the 48-year period between 1963 and 2010. The unbalanced panel
123
data consisted of 1,934 annual observations; certain variables were
missing for some countries and some years in the sample.1 The
countries were classified into two groups: AEC (with the Human
Development Index (HDI)2 in 2010 of 0.86 or greater) and LAEC.
The list of countries included in the sample is presented in
Tables 1 and 2.
The dependent variable was the MPC ownership ratio, and it
was derived by dividing the total number of motorcycles (includ-
ing motorcycles and mopeds) by the total number of passenger
cars for a particular country and the year. The data on motorcycle
ownership and passenger car ownership were derived primarily
from various editions of the International Road Federation (IRF)
World Road Statistic annual yearbooks. Supplement data were
derived from Asean Japan Transport Partnership Information
Centre, United Nations Economic Commission for Europe
Transport Division Database and International Road and Traffic
Accident Database.
The per capita Real Gross Domestic Product (GDP) (Int $ 2005
constant prices: Chain series) was used as a key variable and uti-
lized as a proxy for income. The per capita GDP contributes to
the change in the MPC ownership ratio because it determines the
level of road users’ affordability in purchasing motor vehicles.
The per capita GDP was obtained from the Penn World
Table version 7.0 (Heston et al., 2011). Data on the urban popula-
tion percentage, which was used to examine the impact of urban
population density on the MPC ownership ratio, were obtained
from the World Development Indicator (WDI, 2009). The consumer
price index (CPI)3, a measure for inflation, was used as a proxy for
purchasing power with regard to vehicle purchases. This data is
drawn from the WDI and it is only available for the year 1960
onwards.
The total road length per thousand population was used as
another explanatory variable to determine the MPC ownership
ratio. Previous studies have frequently used the total road length
per thousand population to explain travel patterns and vehicle
ownership (Bento et al., 2005; Ingram and Liu, 1999; Riley,
2002). Data on the this variable were obtained from the WDI and
the IRF databases.
3. Methodology
The use of panel data regression methods has become
increasingly popular with greater availability of panel data for
cross-country data sets. Empirical studies have found that panel
data are generally more informative, which offers greater variabil-
ity, less linearity between variables, and provides more efficient
estimates (Elhorst, 2010; Greene, 2003; Hsiao, 2003). Because the
data used in this study is a panel data set, we used the panel linear
regression with exogenous covariates.
A major concern of using panel data in regression analysis is the
unobserved heterogeneity. Heterogeneity bias refers to the con-
founding effects of unmeasured time-invariant variables, which
are omitted from regression models. Ignoring these effects could
lead to inconsistent estimates of model parameters (Hsiao, 2003).
Econometrically, the use of a fixed effects or random effects
model could control for heterogeneity and provide consistent
and efficient estimates of model parameters in the presence of
1
The study by Shao et al. (2011) indicates that it is common to have unbalanced
panel data for two main reasons. First, the sampling design is not balanced. Second,
some data are missing, although the original design is balanced.
2
The HDI, published annually by the United Nations, measures per capita income,
life expectancy and educational achievement.
3
According to the WDI, CPI reflects changes in the cost to the average consumer of
acquiring a basket of goods and services that may be fixed or changed at specified
intervals, such as yearly.
124 T.H. Law et al. / Journal of Transport Geography 46 (2015) 122–128

Table 1 the per capita GDP showed a negative relationship over most of
The list of Advanced Economic Countries (AEC) and the available years of data. the sample range. This implies that in AEC, increases in per capita
No. Country Observation Period GDP lead to decreases in the MPC ownership ratio. Tables 3 and 4
From To present the descriptive statistics for the explanatory variables used
in this study for AEC and LAEC, respectively. As can be seen, the
1 Australia 30 1963 2004
2 Austria 45 1963 2010
MPC ownership ratio of AEC is lower than that of LAEC. This sug-
3 Belgium 47 1964 2010 gests that economic development may be an important determi-
4 Canada 30 1969 2009 nant of the MPC ownership ratio.
5 Cyprus 41 1963 2010 Table 5 presents a pair-wise correlation matrix between various
6 Czech Rep 18 1993 2010
factors and the MPC ownership ratio. As shown in Table 5, there is
7 Denmark 30 1977 2008
8 Estonia 15 1996 2010 a high correlation between the natural logarithm of per capita GDP
9 Finland 47 1963 2010 and the urban population percentage (0.8592). However, the
10 France 38 1963 2008 results shown in Model F of Table 6 indicate that excluding the nat-
11 Germany 15 1991 2010
ural logarithm of per capita GDP does not significantly affect the
12 Greece 35 1971 2008
13 Hong Kong 14 1981 1994
coefficient for the urban population percentage, indicating that
14 Iceland 46 1963 2008 multicollinearity is not a problem in the estimations.
15 Ireland 39 1965 2010 Table 6 presents the regression results. The explanatory vari-
16 Israel 22 1987 2008 ables used in the analysis are the per capita GDP, the CPI, the urban
17 Italy 40 1963 2010
population percentage and the total road length per thousand
18 Japan 48 1963 2010
19 Korea Rep 33 1971 2008
20 Luxembourg 27 1975 2009
21 Malta 10 1998 2009
Table 2
22 Netherlands 41 1963 2010
The list of LAEC and the available years of data.
23 New Zealand 41 1963 2008
24 Norway 46 1963 2010 No. Country Observation Period
25 Portugal 17 1963 2003
26 Slovakia 15 1995 2010 From To
27 Slovenia 19 1992 2010 1 Bahrain 21 1987 2008
28 Spain 41 1967 2010 2 Bangladesh 9 1990 2003
29 Sweden 33 1973 2009 3 Benin 5 1992 1996
30 Switzerland 48 1963 2010 4 Botswana 18 1975 2005
31 United Kingdom 36 1975 2010 5 Brazil 3 2000 2004
32 United State 40 1970 2010 6 Brunei 16 1991 2010
7 Bulgaria 20 1985 2010
8 Cambodia 6 2005 2010
9 Cameron 12 1968 2006
heterogeneity (Bevan and Danbolt, 2004; Greene, 2003;
10 China 15 1990 2008
Wooldridge, 2002). The fixed effects model assumes that the 11 Colombia 16 1983 2008
country-specific intercept is correlated with the explanatory vari- 12 Costa Rica 20 1984 2008
ables, while the random effects model assumes that the 13 Cote d’Ivoire 10 1966 2007
14 Ecuador 14 1984 2007
country-specific intercept is part of the error term and uncorre-
15 Egypt 17 1982 2008
lated with the explanatory variables (Greene, 2003). 16 Ethiopia 23 1973 2007
The Hausman test was used to determine whether the random 17 Hungary 29 1972 2010
effects model or fixed effects model was more appropriate under 18 India 37 1965 2008
the null hypothesis that the country-specific intercept was uncor- 19 Indonesia 31 1966 2010
20 Jordan 6 2003 2008
related with other explanatory variables in the model. With a cor-
21 Kenya 28 1966 2008
relation (the null hypothesis is rejected), the more appropriate 22 Laos 12 1990 2002
model would be the fixed effects (Baltagi, 2005; Greene, 2003). 23 Latvia 14 1993 2010
The Hausman test indicated that the fixed effects model was more 24 Lithuania 16 1995 2010
25 Malawi 3 1980 1982
appropriate than the random effects model in all models, as shown
26 Malaysia 46 1963 2010
in Table 6. The fixed effects panel linear regression model can be 27 Mauritius 34 1966 2008
written as, 28 Mexico 10 1990 2008
29 Mongolia 10 1994 2003
2
lnðMPCi;t Þ ¼ ai þ x1 lnðGDPi;t Þ þ x2 ðlnðGDPi;t ÞÞ þ uyear i;t 30 Morocco 30 1969 2007
31 Niger 3 2006 2008
þ b0 xi;t þ ei;t ð1Þ 32 Nigeria 11 1973 1996
33 Pakistan 38 1963 2008
where the sub-index i denotes country, t denotes period of time; ln 34 Panama 15 1977 2008
denotes natural logarithms; ai is the country-specific intercept 35 Peru 5 2000 2004
(fixed effects); u, x and b are the model parameters; e is error term; 36 Philippines 18 1981 2002
year is the time trend; x are other explanatory variables. 37 Poland 41 1970 2010
38 Romania 14 1990 2004
39 Senegal 18 1968 2006
4. Results and discussion 40 South Africa 26 1967 2001
41 Sri Lanka 20 1969 2003
42 Swaziland 14 1987 2003
Fig. 1 shows the relationship between the MPC ownership ratio
43 Syria 16 1963 2008
and the per capita GDP. As shown in the figure, for LAEC, the MPC 44 Thailand 35 1967 2010
ownership ratio initially increased and then decreased as income 45 Togo 20 1966 2007
rose. This implies that there is an inverted U-shaped relationship 46 Tunisia 11 1983 2008
47 Turkey 41 1966 2010
between the MPC ownership ratio and the per capita GDP for
48 Vietnam 10 2000 2009
LAEC. On the other hand, for AEC, the MPC ownership ratio and
 T.H. Law et al. / Journal of Transport Geography 46 (2015) 122–128 125

ln (Motorcycles/Passenger Cars)
2

0
5 6 7 8 9 10 11 12

-2

-4

-6

-8 ln (GDP)
LAEC AEC

Fig. 1. The relationship between the motorcycle to passenger car (MPC) ownership ratio and the per capita GDP.

Table 3
Descriptive statistic for Advanced Economic Countries (AEC).

Variable Unit Obs. Mean Std. Dev. Min Max

MPC Motorcycle to passenger car ownership ratio 1047 0.206 0.349 0.001 5.614
GDP Per capita Real GDP (Int $ 2005 constant prices: Chain series) 1047 23,579 9690 3018 73,243
CPI Consumer price index 1047 54.36 30.23 0.06 105.69
URB Urban population (percentage) 1047 72.51 13.69 36.14 99.91
ROAD Road density (km per 1000 population) 1047 16.74 13.65 0.21 82.86

Table 4
Descriptive statistic for Less Advanced Economic Countries (LAEC).

Variable Unit Obs. Mean Std. Dev. Min Max

MPC Motorcycle to passenger car ownership ratio 887 1.884 7.483 0.002 77.012
GDP Per capita Real GDP (Int $ 2005 constant prices: Chain series) 887 5619 7203 349 54,281
CPI Consumer price index (2005 = 100) 887 38.19 31.78 0.0001 101.11
URB Urban population (percentage) 887 44.23 19.56 8.94 88.52
ROAD Road density (km per 1000 population) 887 4.94 5.34 0.39 33.09

Table 5 of Models A to E in Table 6 confirm a statistically significant

Correlation matrix. inverted U-shaped relationship between the MPC ownership ratio
ln(MPC) ln(GDP) ln(CPI) URB ln(ROAD) Year and the per capita GDP. Generally, the MPC ownership ratio
increased with income at a lower level and decreased with income
ln(MPC) 1.0000
ln(GDP) 0.4701 1.0000
at a higher level. The results of Models A and B indicated that the
ln(CPI) 0.1526 0.3226 1.0000 turning point for the MPC ownership ratio was in the range of
URB 0.5501 0.8592a 0.3041 1.0000 US$4895 to US$6060.4
ln(ROAD) 0.4655 0.6915 0.1408 0.5856 1.0000 Model G includes only data after year 1993 to examine the
Year 0.1369 0.2235 0.5414 0.2391 0.0512 1.0000
effect of per capita GDP on the MPC ownership ratio at a higher
a
Correlation value > 0.8. per capita GDP level. The average GDP per capita between 1993
and 2010 for LAEC and AEC was USD$7812 and USD$28,975,
respectively. The results showed a significant U-shaped relation-
population. The natural logarithm of the explanatory variables are ship between the MPC ownership ratio and the per capita GDP.
used to minimize heteroscedasticity and also allow easier interpre- Specifically, the MPC ownership ratio declined with per capita
tation of the relative elasticity value of the estimates. A quadratic GDP at a lower per capita GDP level and increased with per capita
term was added to the per capita GDP and the urban population GDP at a higher per capita GDP level.5 The turning point was
percentage to account for the quadratic relationship between these reached at US$9215. This finding may be attributed to people with
two variables and the MPC ownership ratio. The CPI and the total a higher income level using motorcycles in recreational and leisure
length per thousand population interacted with the per capita activities, especially in AEC (Jamson and Chorlton, 2009).
GDP. This allows us to examine the effect of total road length per The estimated results showed that the coefficient for the CPI was
thousand population and CPI on the MPC ownership ratio as positive and statistically significant, whereas the coefficient on its
income increased. A time trend is included to capture the effects interaction with the per capita GDP was estimated to be negative.
of other exogenous time dependent variables.
Model A only includes linear and quadratic terms of per capita
4
GDP to investigate the relationship between the MPC ownership The turning point is calculated by, exp(x1/(2.x2)) where x1 is the coefficient on
the log of GDP and x2 is the coefficient on the log GDP, whole quantity squared.
ratio and the per capita GDP. Models B to F include the influence 5
Model G in Table 6 was also estimated based on data after year 1984, and similar
of additional explanatory variables on this relationship. The results results were found (not reported).
126 T.H. Law et al. / Journal of Transport Geography 46 (2015) 122–128

Table 6
Analysis results.

Model A Model B Model C Model D Model E Model F Model G

Variables
ln(GDP) 6.151*** 6.027*** 5.613*** 3.557*** 5.233*** 1.917**
(ln(GDP))2 0.362*** 0.346*** 0.318*** 0.237*** 0.297*** 0.105**
ln(CPI) 0.231* 0.259* 0.681*** 1.389***
ln(CPI) ⁄ ln(GDP) 0.034** 0.037** 0.085*** 0.164***
ln(ROAD) 1.035*** 0. 602* 1.365***
ln(ROAD) ⁄ ln(GDP) 0.093** 0.051 0.131***
URB 0.069*** 0.051***
URB2 0.0004*** 0.0002***
Year 0.0052** 0.0012 0.0028 0.0029
Constant 27.264*** 17. 203*** 28.284*** 18.451*** 35.834*** 1.847*** 6.703*
No. Obs. 1934 1934 1934 1934 1934 1934 821
No. Groups 80 80 80 80 80 80 79
Overall model significance 130*** 131*** 132*** 126*** 109*** 107*** 205***
Hausman test (Chi-square) 27.29*** 32.34*** 19.98*** 18.89** 18.36** 24.65*** 18.99***

Model G includes only data after year 1993.

*
Significant at 10%.
**
Significant at 5%.
***
Significant at 1%.

Several studies have shown that inflation increases the price of
goods and services, thereby eroding purchasing power (Manzoor
et al., 2011; Molana, 1990). Therefore, under high inflation circum-
stances, commuters tend to opt for vehicles with lower purchase
and running costs, such as motorcycles and mopeds, and this leads
to an increase in the MPC ownership ratio at a lower income level.
Nonetheless, the effect of CPI on the MPC ownership ratio declined
as income continued to rise. This is probably because cars become a
necessity 6and are more affordable at a higher income level.
The quadratic coefficient for urbanisation was positive and sig-
nificant in Models E and F, pointing to a convex relationship
between urbanisation and the MPC ownership ratio. At a lower
urban population density level, the scale of economic activities is
small and less varied in scope than at a higher urban population
density. Therefore, there is less of a demand for long-distance tra-
vel, and motorcycles are preferred to passenger cars may be
because they are more suited to short-distance travel. However,
as the urban population density and urban area increase,
long-distance travel by passenger car is expected to increase. On
the other hand, a higher level of urban population growth may give
rise to traffic congestion, as well as strong competition for parking
spaces. Karathodorou et al. (2010) and Ng et al. (2010) indicated
that traffic congestion and limited parking areas discouraged
potential passenger car owners. Nishitateno and Burke (2014) indi-
cated that some commuters may choose to ride a motorcycle to
avoid traffic congestion and limited parking bays. As a result, the
MPC ownership ratio increases.
Models D and F showed that the total road length per thousand
population had a positive and statistically significant effect on the
MPC ownership ratio, whereas the coefficient for its interaction
with per capita GDP was estimated to be negative. These results
suggest that the MPC ownership ratio increases with total road
length per thousand population at a low income level. However,
this effect declined with per capita GDP growth over time. A possi-
ble explanation for this finding is that road network development
was mainly limited to urban areas and low-performance roads
(such as unpaved and lower number of lanes) in the early stages
of road network development. At this development phase, a higher
road density was associated with enhanced local accessibility.
Motorcycle ownership can be expected to rise with road density
at this development stage. However, as an economy grows, more
6
Individuals who own cars are more probable to be employed (Ong, 1996; Holzer
et al., 1994).
extensive high-performance road networks (such as expressway)
are built to connect the large cities. At this stage of road develop-
ment, motorcycles can be expected to be less preferable than pas-
senger cars, possibly because motorcycles are less safe and less
comfortable than cars for longer-distance travel. Consequently,
the MPC ownership ratio declines.
Model B indicated that the time trend variable was negative and
significant, implying that some unmeasured effects caused a
downward trend in the MPC ownership ratio over time.
However, the inclusion of the CPI and its interaction with the per
capita GDP eliminated the significance of the time trend variable
in Models C, D and E, suggesting that these variables played a role
in reducing MPC ownership ratio.
5. Conclusions
This study focused on the effect of economic growth on the MPC
ownership ratio. The estimated results pointed to an inverted
U-shaped relationship between the MPC ownership ratio and the
per capita GDP. The evidence indicated that urbanisation, the total
road length per thousand population, and the CPI were the underly-
ing factors that contributed to this relationship. In view of the rela-
tionship, it could be assumed that at a low income level, motorcycle
ownership is popular because the scale of the economic activities is
relatively low and simple. Moreover, cars are less affordable than
motorcycles at a low income level. As the level of income increases,
purchasing power also rises. At this stage, more people likely switch
from motorcycles to passenger cars probably for their prestige, con-
venience and safety. At a higher income level, more resources are
also available for the government to invest in improvements of road
network infrastructures. Advances in road network infrastructures
are likely to lead to increased passenger car ownership. The latter
is likely probably due to passenger cars being safer than motorcy-
cles and more suitable for long-distance travel.
A limitation of this study is that the models omitted other trans-
portation modes, such as public transportation, bicycling and walk-
ing. Earlier studies have consistently pointed to a higher demand
for public transportation in densely populated areas (Lam and
Tam, 2002; Clark, 2007, 2009; Khan and Willumsen, 1986). This is
probably due to the higher level of congestion, greater parking con-
straints and costs, and more organized public transportation sys-
tems in these areas (Dargay and Gately, 1999; De Jong and Van de
Riet, 2008; Schwanen et al., 2004). Previous studies also showed
that public transportation tended to impose lower risks on
 T.H. Law et al. / Journal of Transport Geography 46 (2015) 122–128
passengers (Brenac and Clabaux, 2005; Elvik and Bjørnskau, 2005;
Evans, 1994). In this study, we assumed that motorcycles and pas-
senger cars were the two major private transportation modes, with
commuters having to choose one mode of travel. In reality, they
could select other private transportation modes, such as bicycling
and walking, to meet their mobility needs. The factors that influ-
ence the use of these private transportation modes may differ from
those that influence the use of motorcycles and passenger cars.
Therefore, further research is needed to examine the modal shift
from motorcycles to other transportation modes and to determine
what factors contribute to this change.
The key finding of this study is that there are complex pathways
from economic growth to the change in the MPC ownership ratio,
such as urbanisation, advances in road network infrastructures,
and vehicle purchasing power. However, there is always a time
lag between economic growth and the MPC ownership ratio. That
is, economic growth may be rapid but the effects of transport poli-
cies and measures may take longer time to be felt, particularly for
some newly-developed cities in LAEC. This may have implications
on transport policies and strategies. Future research to address this
issue is warranted and should be pursued.
Many of the LAEC in the sample have per capita GDP below the
turning point of the inverted U-curve for the MPC ownership ratio.
This implies that these countries are still at a stage where further
economic growth will result in an increase in motorcycle owner-
ship. As reported by Law et al. (2009), growth in income at lower
income levels appeared to be associated with a rise in motorcycle
deaths, with the increase partly explained by the large number of
motorcycles and partly by the high degree of mobility of the motor-
cyclists. To enhance motorcycle safety, these countries could focus
on the provision of specific injury-prevention countermeasures.
Such measures could include improving public transportation ser-
vices and simultaneously encouraging motorcyclists to use public
transportation instead of riding motorcycles, as well as designing
roads with exclusive lanes for motorcycles and increasing the strin-
gency of motorcycle helmet laws and enforcement.
As indicated in the present study, the MPC ownership ratio
declined once the per capita GDP level exceeded a threshold level.
This result suggests that developed countries, particularly those
with increased urbanisation and a higher road network density,
should devote resources to accommodating the increased demand
for passenger cars, without causing excessive congestion. The
empirical evidence found in this study also showed that the reduc-
tion in the MPC ownership ratio after the first turning point was
not permanent, with a second turning point occurring at a higher
per capita GDP level. Jamson and Chorlton (2009) indicated that
at higher income levels, more mature motorcycle riders take up
motorcycling as a leisure pursuit rather than as a means of trans-
port. Therefore, more emphasis should be placed on motorcycling
training specifically for leisure motorcycle riders.
Acknowledgement
This research was funded by Universiti Putra Malaysia Research
University Grant (Grant No. 05-05-10-1108RU). The funding
source played absolutely no role in the study design nor any collec-
tion, analysis, or interpretation of the data.
References
Abdul Manan, M.M., Várhelyi, A., 2012. Motorcycle fatalities in Malaysia. IATSS Res.
36 (1), 30–39.
Baltagi, B.H., 2005. Econometric Analysis of Panel Data. John Wiley and Sons Inc.,
New York.
Bento, A.M., Cropper, M.L., Mobarak, A.M., Vinha, K., 2005. The effects of urban
spatial structure on travel demand in the United States. Rev. Econ. Stat. 87 (3),
466–478.
127
Bevan, A.A., Danbolt, J., 2004. Testing for inconsistencies in the estimation of UK
capital structure determinants. Appl. Financ. Econ. 14 (1), 55–66.
Brenac, T., Clabaux, N., 2005. The indirect involvement of buses in traffic accident
processes. Saf. Sci. 43 (10), 835–843.
Button, K., Ngoe, N., Hine, J., 1993. Modelling vehicle ownership and use in low
income countries. J. Transport Econ. Policy 27 (1), 57–67.
Cao, X., Huang, X., 2013. City-level determinants of private car ownership in China.
Asian Geogr. 30 (1), 37–53.
Chiou, Y.C., Wen, C.H., Tsai, S.H., Wang, W.Y., 2009. Integrated modelling of
car/motorcycle ownership, type and usage for estimating energy consumption
and emissions. Transport. Res. Part A: Policy Pract. 43 (7), 665–684.
Clark, S., 2007. Estimating local car ownership models. J. Transport Geogr. 15 (3),
184–197.
Clark, S., 2009. Charactering and predicting car ownership using rough sets.
Transport. Res. Part C: Emerg. Technol. 17 (4), 381–393.
Dao, N.X., Duc, N.H., 2005. Study on the ownership of motorcycle in Vietnam. In:
Proceeding the 6th, International Conference of Eastern Asia Society for
Transportation Studies, Bangkok.
Dargay, J., Gately, D., 1999. Income’s effect on car and vehicle ownership,
worldwide: 1960–2015. Transport. Res. Part A: Policy Pract. 33 (2), 101–138.
Dargay, J., Hanly, M., 2007. Volatility of car ownership, commuting mode and time
in the UK. Transport. Res. Part A: Policy Pract. 41 (10), 934–948.
Dargay, J.M., Vythoulkas, P.C., 1999. Estimation of a dynamic car ownership model:
a pseudo-panel approach. J. Transport Econ. Policy 33 (3), 287–302.
De Jong, G.C., Van de Riet, O., 2008. The driving factors of passenger transport. Eur. J.
Transport Infrastruct. Res. 8 (3), 227–250.
De Jong, G., Fox, J., Pieters, M., Daly, A.J., Smit, R., 2004. Comparison of car ownership
models. Transport Rev. 24 (4), 379–408.
Delbosc, A., 2013. Household composition and within-household car saturation in
Melbourne. Transport Policy 25, 94–100.
Elhorst, J.P., 2010. Spatial panel data models. In: Fischer, M.M., Getis, A. (Eds.),
Handbook of Applied Spatial Analysis: Software Tools, Methods and
Applications. Springer, Berlin, pp. 377–407.
Elvik, R., Bjørnskau, T., 2005. How accurately does the public perceive differences in
transport risks?: an exploratory analysis of scales representing perceived risk.
Accid. Anal. Prevent. 37 (6), 1005–1011.
Evans, A.W., 1994. Evaluating public transport and road safety measures. Accid.
Anal. Prevent. 26 (4), 411–428.
Greene, W.H., 2003. Econometric Analysis. Prentice Education, Upper Saddle River,
NJ.
Hess, D.B., Ong, P.M., 2002. Traditional neighbourhoods and automobile ownership.
Transport. Res. Rec. 1805, 35–44.
Heston, A., Summers, R., Aten, B., 2011. Penn world table version 7.0. Centre for
International Comparisons of Production, Income and Prices. University of
Pennsylvania.
Holzer, H.J., Ihlanfeldt, K.R., Sjoquist, D.L., 1994. Work, search and travel among
white and black youth. J. Urban Econ. 35, 320–345.
Hsiao, C., 2003. Analysis of Panel Data, second ed. Cambridge University Press,
Cambridge, UK.
Hsu, T.P., 2005. Comparative study on motorcycle ownership forecasting model of
Asian Countries-Taiwan, Malaysia and Vietnam. In: Proceedings the 6th
International Conference of Eastern Asia Society for Transpiration Studies,
Bangkok.
Ingram, G.K., Liu, Z., 1999. Determinants of motorization and road provision. World
Bank Policy Research Working Paper No. 2042. The World Bank, Washington,
D.C.
Jamson, S., Chorlton, K., 2009. The changing nature of motorcycling: patterns of use
and rider characteristics. Transport. Res. Part F: Traffic Psychol. Behav. 12 (4),
335–346.
Karathodorou, N., Graham, D.J., Noland, R.B., 2010. Estimating the effect of urban
density on fuel demand. Energy Econ. 32 (1), 86–92.
Khan, A., Willumsen, L., 1986. Modelling car ownership and use in developing
countries. Traffic Eng. Control 27 (11), 554–560.
Lai, W.T., Lu, J.L., Chiang, Y.S., 2006. The study of multi-vehicle household mix
demand model on auto–motorcycle ownership and use in Taiwan area.
Transport. Plann. J. 35 (3), 309–336.
Lam, W., Tam, M., 2002. Reliability of territory-wide car ownership estimates in
Hong Kong. J. Transport Geogr. 10 (1), 51–60.
Law, T.H., Noland, R.B., Evans, A.W., 2009. Factors associated with the relationship
between motorcycle deaths and economic growth. Accid. Anal. Prevent. 41 (2),
234–240.
Manzoor, M.M., Siddiqui, A.A., Sattar, A., Fahim, M., Rasheed, S., 2011. Impact of
inflation on household consumption – a case of Pakistan. Int. Res. J. Finance
Econ. 68, 161–167.
Molana, H., 1990. A note on the effect of inflation on consumers’ expenditure.
Empirical Econ. 15 (1), 1–15.
Nagai, Y., Fukuda, A., Okada, Y., Hashino, Y., 2003. Two-wheeled vehicle ownership
trends and issues in the Asian region. J. Eastern Asia Soc. Transport. Stud. 5,
135–146.
Ng, W.S., Schipper, L., Chen, Y., 2010. China motorization trends: new directions for
crowded cities. J. Transport Land Use 3 (3), 5–25.
National Highway Transportation Safety Administration (NHTSA), 2008. Traffic
Safety Facts 2007 Data: Motorcycles (Report No. DOT HS 810 990). NHTSA,
National Centre for Statistics and Analysis, Washington, D.C.
Nishitateno, S., Burke, P.J., 2014. The motorcycle Kuznets curve. J. Transport Geogr.
36, 116–123.
128 T.H. Law et al. / Journal of Transport Geography 46 (2015) 122–128
Nolan, A., 2010. A dynamic analysis of household car ownership. Transport. Res. Part
A: Policy Pract. 44 (6), 446–455.
Ong, P., 1996. Work and automobile ownership among welfare recipients. Soc.
Work Res. 30 (4), 255–262.
Pongthanaisawan, J., Sorapipatana, C., 2010. Relationship between level of economic
development and motorcycle and car ownerships and their impacts on fuel
consumption and greenhouse gas emission in Thailand. Renew. Sustain. Energy
Rev. 14 (9), 2966–2975.
Preusser, D.F., Williams, A.F., Ulmer, R.G., 1995. Analysis of fatal motorcycle crashes:
crash typing. Accid. Anal. Prevent. 27 (6), 845–851.
Radin Umar, R.S., Mackay, G.M., Hills, B.L., 1995. Preliminary analysis of motorcycle
accidents: short-term impacts of the running headlights campaign and
regulation in Malaysia. J. Traffic Med. 23, 17–28.
RAND Europe and WSP, 2004. Motorcycles and Congestion: The Effect of Modal Split
– Full Summary, Department for Transport (with support from Transport for
London).
Ranney, M.L., Mello, M.J., Baird, J.B., Chai, P.R., Clark, M.A., 2010. Correlates of
motorcycle helmet use among recent graduates of a motorcycle training course.
Accid. Anal. Prevent. 42 (6), 2057–2062.
Riley, K., 2002. Motor vehicles in China: the impact of demographic and economic
changes. Populat. Environ. 23 (5), 479–494.
Romilly, P., Song, H., Liu, X., 1998. Modelling and forecasting car ownership in
Britain: a cointegration and general to specific approach. J. Transport Econ.
Policy 132 (2), 165–185.
Sanko, N., Maesoba, H., Dissanayake, D., Yamamoto, T., Kurauchi, S., 2009. Inter-
temporal analysis of household car and motorcycle ownership behaviours: the
case in Nagoya Metropolitan Area of Japan, 1981–2001. IATSS Res. 33 (2), 39–53.
Schwanen, T., Dijst, M., Dieleman, F.M., 2004. Policies for urban form and their
impact on travel: the Netherlands experience. Urban Stud. 41 (3), 579–603.
Sehatzadeh, B., Noland, R.B., Weiner, M.D., 2011. Walking frequency, cars, dogs and
the built environment. Transport. Res. Part A: Policy Pract. 45 (8), 741–754.
Senbil, M., Zhang, J., Fujiwara, A., 2007. Motorization in Asia – 14 countries and
three metropolitan areas. Int. Assoc. Traffic Saf. Sci. Res. 31 (1), 46–58.
Shao, J., Xiao, Z., Xu, R., 2011. Estimation with unbalanced panel data having
covariate measurement error. J. Stat. Plann. Inference 141 (2), 800–808.
Sharma, B.R., 2008. Road traffic injuries: a major global public health crisis. Public
Health 122 (12), 1399–1406.
Sillaparcharn, P., 2007. Vehicle ownership and trip generation modelling: a case
study of Thailand. J. Int. Assoc. Traffic Saf. Sci. 31 (2), 17–26.
Tanner, J.C., 1978. Long-term forecasting of vehicle ownership and road traffic. J.
Roy. Stat. Soc., Ser. A 141 (1), 13–63.
Tanner, J.S., 1983. International comparisons of car ownership and car usage.
Transport and Road Research Laboratory Report 1070. Department of the
Environment and Transport, Crowthorne, Berkshire.
Tuan, V.A., 2011. Dynamic interactions between private passenger car and
motorcycle ownership in Asia: a cross-country analysis. J. Eastern Asia Soc.
Transport. Stud. 9, 541–556.
Tuan, V.A., Shimizu, T., 2005. Modelling of household motorcycle ownership
behaviour in Hanoi city. J. Eastern Asia Soc. Transport. Stud. 6, 1751–1765.
World Development Indicators (WDI), 2009. World Bank World Development
Indicators January 2009, ESDS International, (MIMAS) University of Manchester.
Whelan, G., 2001. Methodological advances in modelling and forecasting car
ownership in Great Britain. In: Paper for European Transport Conference 2001,
PTRC, Cambridge.
Whelan, G., 2007. Modelling car ownership in Great Britain. Transport. Res. Part A:
Policy Pract. 41 (3), 205–219.
Whelan, G., Wardman, M., Daly, A., 2000. Is there a limit to car ownership growth?
An exploration of household saturation levels using two novel approaches. In:
Paper presented at European Transport Conference 2000, PTRC, Cambridge.
Wong, K.I., 2013. An analysis of car and motorcycle ownerships in Macao. Int. J.
Sustain. Transport. 7 (3), 204–225.
Wooldridge, J.M., 2002. Econometric Analysis of Cross Section and Panel Data. MIT
Press, Cambridge, MA.
Yamamoto, T., 2009. Comparative analysis of household car, motorcycle and bicycle
ownership between Osaka metropolitan area, Japan and Kuala Lumpur,
Malaysia. Transportation 36 (3), 351–366.