Content uploaded by Kurt Kratena
Author content
All content in this area was uploaded by Kurt Kratena on Jul 09, 2015
Content may be subject to copyright.
Content uploaded by Kurt Kratena
Author content
All content in this area was uploaded by Kurt Kratena on Jul 09, 2015
Content may be subject to copyright.
E STUDIOS DE ECONOMÍA APLICADA VOL. 21 - 2, 2 0 0 3. P ÁGS. 243-258
Modelling Consumer Transport Demand and Sustainable
Development
KRATENA, K(*) y WÜGER, M.
*
Austrian Institute of Economic Research (WIFO).
PO Box 91, 1103 Vienna Tel +43 (0) 1 7982601-0 E-mail: kurt.kratena@wifo.ac.at.
RESUMEN
El presente artículo se centra en la modelización económica y el análisis empírico de estructuras
sostenibles del consumo privado, tratando de extender los modelos económicos convencionales de
consumo. El punto de partida para el análisis de consumo sontenible de energía para el transporte es el
concepto de funciones de producción de los hogares. El punto principal del análisis son los servicios de
consumo derivados de una combinación de stocks (sistema de transportes) y flujos (principalmente,
energía). Los patrones de consumo sostenible pueden alcanzarse mediante una sustitución de flujos
por stocks (por ejemplo, mejoras en la eficiencia energética del sistema de transportes). Los dos facto-
res esenciales en el contexto del consumo sostenible son, por un lado, los cambios en la demanda de
los servicios de consumo deseados y, por otro, la estructura de la combinación entre flujos y stocks
necesarios para la provisión de dichos servicios.
Palabras clave: desarrollo sotenible, consumo, modelización económica.
ABSTRACT
The paper focuses on economic modelling and empirical analysis of sustainable structures in private
consumption and strives to extend conventional economic consumption models. Starting point for the
model analysis of sustainable consumption of energy for transport purposes was the household production
function concept. The focal point of the analysis is consumer services derived from a combination of
stocks (transport systems) and flows (mainly energy). Sustainable consumption patterns can arise,
when within service demand production substitution of flows by stocks (e.g. improvements in energy
efficiency of transport systems) takes place. Two essential factors are crucial in the context of sustainable
consumption: the demand shifts concerning the consumer services desired, and the composition of the
stock-flow mix necessary for the service provision.
Keywords: Sustainable Development, Consumption, Economic Modelling
Clasificación JEL: Q01, D11.
1. INTRODUCTION
In the past decades, environmental debate tended to concentrate largely on the
negative environmental impacts of production processes. In 1992, the Rio Earth Summit
Artículo recibido en abril de 2003. Aceptado en junio de 2003.
Articulo-04.p65 22/07/03, 9:33243
244
K. Kratena y M. Wüger
Estudios de Economía Aplicada, 2003: 243-258 • Vol. 21-2
led to a more holistic approach, with the concept of sustainable development becoming
an accepted part of scientific and political discourse. Given the importance of private
consumption in the economy, consumer behaviour and lifestyles are increasingly
recognised as determining factors for sustainable development. Consumption structures
influence production processes and involve resource use. Sustainable development can
be achieved by a “decoupling” of resource use and consumer expenditure. If consumption
patterns are to be changed that might imply consequences for welfare, as resource
intensive goods also might contribute to consumer´s welfare. Another option therefore
is technological changes, so that “decoupling” might be achieved without important
welfare implications. A third approach that will be followed here is that resource
intensive goods must be seen as inputs into a consumer´s production process, where
welfare relevant services (e.g. mobility) are produced. In such a framework “decoupling”
can be achieved by different changes in service demand and in the production process
of these services.
Economic research on the causes and consequences of material intensive consumer
behaviour and its negative impacts on the environment is characterised by a great
variety of theoretical and methodological approaches that go beyond neo-classical
consumption theory. The research on sustainable consumption strives to integrate various
disciplines in order to depict the driving forces of consumer behaviour and to derive
policy instruments aiming at changing consumer behaviour. One line of the literature
puts the emphasis on the criticism of the neo-classical “homo oeconomicus”
(Siebenhüner, 2000, Duchin, 1998). Intensive research has been carried out about the
driving forces for material intensive consumption, focusing on economic, socio-
psychological and socio-technical explanations (Røpke, 1999, 2001, Wilk, 2001, Brown
and Cameron, 2000, Wilhite et al., 1996, Shove and Warde, 1997). Another line of
analysis has undertaken modelling of consumer behaviour within economic-ecological
models (Gintis, 2000, Bossel, 2000, Jager et al., 2000, 2001).
The integration of the above mentioned research into economic modelling is missing
to a large extent. In the present paper we attempt to carry out a modelling exercise in
the framework of a consumer´s production process, where welfare relevant services
are produced. Total (market) demand shall additionally be described as an aggregation
over different groups of household with specific consumption patterns. As in Wenke
(1993) utility in consumption is not solely determined by the quantity of goods consumed,
but also by additional components, which are related to a specific behaviour, like
sustainable consumption patterns. In this way, shifts in preference structures can be
depicted. As the stock-flow relationship can be seen as crucial, an approach of integrated
modelling for durable and non-durable goods demand as in Conrad and Schröder (1991)
would be necessary.
Conrad and Schröder (1991) stay within a traditional neo-classical model where
cost-minimisation determines capital stock accumulation. Demand shifts and changes
in consumer behaviour towards more sustainable consumption patterns come in such
Articulo-04.p65 22/07/03, 9:33244
245MODELLING CONSUMER TRANSPORT DEMAND AND SUSTAINABLE DEVELOPMENT
Estudios de Economía Aplicada, 2003: 243-258 • Vol. 21-2
an approach only via changes in prices of resources and/or capital. The most impor-
tan and promising approach for modelling stock-flow relationships in consumer
demand is the household production function, derived originally by Becker (1965).
According to Deaton and Muellbauer (1980): “the household production approach is
not merely a clever or elegant way for looking at household decisions but the only
appropriate way.”
The work presented here set as its target the economic modelling and quantification
of changes in consumption behaviour in the category transport. The aim was to evaluate
options of economic policy as well as potential changes in lifestyles and their respecti-
ve impacts on energy flows. Here it is intended to move well beyond the standard
models of consumption by incorporating non-economic factors into empirical analysis
and modelling. The emphasis of the analysis is put on consumption services and the
depiction of the interaction between stocks (capital) and flows (energy and other
materials). Further, the analysis attempts to illustrate the impact of shifts in demand
resulting from changes in consumption patterns.
2. THE HOUSEHOLD PRODUCTION FUNCTION
The household production approach establishes a link between the theory of
consumption and the theory of the firm (Roth, 1998). Decision making is seen to occur
in two stages, with stocks and technologies playing important roles. Particular attention
is devoted to the treatment of non-market activity, time, and the state of consumer
knowledge (human capital). This approach focuses specifically on the conversion of
purchasable goods into so-called “commodities”. While in traditional economic theory
consumption analysis focuses on the demand for goods, in the theory of household
production it is “commodities” which are demanded and provide utility. These
incorporate additional characteristics or benefits (comfort, entertainment, nourishment,
etc.), and are produced subject to the prevailing consumption technology constraints,
which are included in the household production function.
The original idea of the household production function put forward by Lancaster,
1966 has been taken up by various authors to show the differences to traditional
consumption theory (Stigler and Becker, 1977). A very interesting application to energy
consumption including investment decisions in energy efficiency can be found in Willett
and Naghshpour, 1987. These studies do not include empirical applications of the
household production function or derivation of explicit demand functions.
Some studies describe household production by use of time inputs from the household
(instead of capital) and derive demand functions, that can be estimated with econometric
methods (e.g. Graham and Green, 1984).
Treating transport or total energy consumption by households as an input into the
production of a service (mobility or room temperature), we assume, that it is ultimately
this service output, that creates utility. In the utility function U therefore we find
Articulo-04.p65 22/07/03, 9:33245
246
K. Kratena y M. Wüger
Estudios de Economía Aplicada, 2003: 243-258 • Vol. 21-2
consumption of non-energy goods, C
NE
as well as these services, SE, produced with
energy inputs, E and capital inputs, K by – for example – a Cobb-Douglas production
function:
U = [C
NE
, SE] (1)
SE = α
0
E
α1
K
α2
(2)
In production theory the usual way to proceed would be to derive conditional factor
demand functions
SE
K
,
SE
E
directly from (2), depending on relative prices for these
inputs, p
E
and p
K
, where the latter should incorporate all elements of the user costs of
capital:
=
k
E
p
p
f
SE
E
(3)
with the partial derivative to the own price f p
E
< 0.
The capital stock K follows an accumulation path in time t (subscript) determined
by household investment I
t
and depreciation rate δ:
K
t
= K
t-1
+ I
t
–
δδ
δδ
δK
t-1
(4)
The investment decision of households therefore has a direct impact on energy
inputs in the production of SE via capital. The model is complemented by adding a
budget restriction of disposable household income with savings S:
YD = p
E
E + p
NE
C
NE
, + S (5)
The strictly neo-classical approach of the household production function (Willett –
Naghspour, 1987) assumes that all household investment has to equal savings:
S = p
I
I (6)
so that optimization also includes the investment decision.
The household than maximizes the utility function (1), where we can insert the
household production function under the restrictions of capital accumulation and
exhaustion of the budget:
Articulo-04.p65 22/07/03, 9:33246
247MODELLING CONSUMER TRANSPORT DEMAND AND SUSTAINABLE DEVELOPMENT
Estudios de Economía Aplicada, 2003: 243-258 • Vol. 21-2
max [(α
0
E
α1
K
α2
), C
NE
] (7)
K = K
*
or K
t
= K
t-1
+ I
t
– δK
t-1
YD – S - p
E
E - p
NE
C
NE
= 0
If we apply the capital accumulation function, then we face a problem of
intertemporal maximization, where we can additionally introduce the restriction (6) of
savings and investment equality. This leads to a household investment function derived
also from utility maximization.
Problem (7) can be solved to derive explicit demand functions for E and C
NE
, if we
assume an explicit utility function. The partial derivative of utility to goods can as far
as energy is concerned be decomposed in this model by the product of the partial
derivative of utility to the services SE (
∂
U/
∂
SE) and the production elasticity of an
input (
∂
SE/
∂
E). As in the traditional consumption model the marginal rate of
substitution calculated from these derivatives must equal relative prices (Willett and
Naghshpour, 1987, p. 253). Therefore it is mainly relative prices that determine
consumption of goods and services. In the purely neoclassical model all households
decisions are solely determined by these relative prices, including the investment decision.
3. ESTIMATION RESULTS FOR PRIVATE TRANSPORT CONSUMPTION
In the following section an empirical model for private consumption is presented.
This model starts from a household production function approach for transport (or
better: mobility) demand and therefore splits up total private consumption into the
categories: transport and non-transport consumption. Although starting from this concept
we deviate from the purely neoclassical model in some important features and present
a more flexible approach with less a priori restrictions following from microeconomic
theory. First of all, it is assumed that no explicit substitution takes place between the
two consumption categories. As in the model outlined above, we use the budget
restriction of the household (5). For any given levels of total consumption, and given
levels of expenditure for transport, non-energy consumption is treated as a residual
element, what actually implies a substitution elasticity of 1 and follows from (5), if
total private consumption in nominal terms (= total expenditure, p
C
C) is given:
p
NE
C
NE
, = p
C
C - p
E
E - p
I
I (8)
The substitution of energy flows by “units” of capital can only be achieved at the
expense of short-term non-energy consumption. The corresponding investment and
capital variable for transport are vehicle stock and public transport infrastructure (roads,
Articulo-04.p65 22/07/03, 9:33247
248
K. Kratena y M. Wüger
Estudios de Economía Aplicada, 2003: 243-258 • Vol. 21-2
public transport networks). The household budget restriction includes only vehicles
in our model, as there is no direct linkage between non-energy consumption and
transport infrastructure capital. This could to some extent be achieved by the adoption
of a public sector budget constraint, such that higher investment in transport
infrastructure might partly be financed by reductions in disposable income (e.g. through
extra taxation or cuts in other areas of public spending).
∆log(C) = C {∆log(YD/p
C
), ECM} (9)
Equation (9) describes total real private consumption C as a function of real
household disposable income, with YD as nominal income and PC as total consumption
deflator, where a dynamic specification with error correction mechanism (ECM) was
chosen. The savings decision is therefore described by this macroeconomic consumption
function and not derived from utility maximization.
Another important part of the model, where we do not follow the purely neoclassical
line is capital accumulation. Capital accumulation is by definition driven by investment
and depreciation as in (4):
K
t
= K
t-1
+ I
t
- δK
t-1
(10)
We assume that capital stock of vehicles also contributes to the utility of the household
and depends on the variable Z and is described by stock adjustment equations, where
a
1
and a
2
are the adjustment terms. The aggregate variable Z includes variables which
exert an influence on capital accumulation, i.e. mainly real disposable income and fuel
prices. The accumulation process of infrastructure investments is not explicitly modelled,
but this capital stock also determines the combinations of capital and energy used to
produce energy services.
∆log K
t
= F {Z
t
, α
1
log K
t-1
, α
2
∆log K
t-1
} (11)
Additionally the impact of prices and taxes on average fuel consumption for the
vehicle fleet were also tested and taken into account in a ‘capital stock-efficiency’
equation. Actually our treatment of capital accumulation is a very flexible form of the
microeconomic investment function, because relative prices play some role. This function
would better fit with an utility function including the stock of capital as welfare relevant:
U = [C
NE
, SE, K] (12)
The vehicle fleet adds to household utility by its existence and not only via the
household production function. Capital accumulation is therefore not only guided by
prices together with energy demand as a factor input, but rather represents a “quasi
fixed” factor for energy demand:
Articulo-04.p65 22/07/03, 9:33248
249MODELLING CONSUMER TRANSPORT DEMAND AND SUSTAINABLE DEVELOPMENT
Estudios de Economía Aplicada, 2003: 243-258 • Vol. 21-2
=
A
POP
,
p
p
,
SE/K
SE/K
F
SE
EN
EN
o
o
p
(13)
Equation (13) incorporates the relation between both capital stocks (public K
O
and
private transport K
p
), the relative price of public transport (p
O
/p
EN
), plus “relative
population density” (=ratio of population in areas surrounding cities to respective city
population, POP/A), this latter serving as a general variable for mobility demand biased
towards passenger car transport. An increase in the SE/K ratio implies that mobility
demand rises in relation to transport infrastructure thereby inducing additional energy
(fuel) demand for transport. A higher relative price of public transport also raises
energy (fuel) demand.
Equations describing the demand for energy services complement the model. The
specification of our model applies to a situation, where these energy service variables
can be approximated by specific variables (passenger kilometres for mobility). Energy
service SE is represented by total mobility, i.e. the sum of passenger kilometres for all
forms of transport. This is determined by macro-economic factors such as real disposable
income, as well as by relative prices, with p
SE
being the national accounts figure for
current transport expenditure (i.e. excluding vehicle purchases). Population density
was also introduced to measure its impact on total mobility demand.
log SE = F {log(YD/p
C
), log (p
SE
/ p
C
), log (POP/A)} (14)
Energy demand (EN) given in national accounting figures represents the average
for various household types, and that these types also vary in terms of their consumption
sustainability. Quantification of these differences in household consumption patterns
is derived from consumer survey data. For this purpose we used a fixed effect panel
regression, which allows us to derive ‘non-explained’ demand patterns. Households
can therefore be ranked in terms of energy expenditure, taking criteria such as population
density, fuel consumption (derived from vehicle tax rates) into account, and the two
types of households, “sustainable” and “normal” are then determined. “Sustainable”
households exhibit values for expenditure on fuel below the median, while for “nor-
mal” households such expenditures are above the median. Total expenses on a macro
level are then given by:
EN = x
1
EN
S
+ (1 - x
1
) EN
N
, (15)
where EN
S
and
EN
N
are averages for the sums of “normal” (N) households and
“sustainable” households consumption. For the initial median cut off, the value of x
1
was exactly 0.5. On implementing demand shifts, and based on the target values set for
EN, the relevant weights x
1
and (1–x
1
) are recalculated, i.e. it is calculated how many
of the “normal” households must change to “sustainable” households in order to reach
a target EN. Any additional effects can be estimated for total consumption, when the
Articulo-04.p65 22/07/03, 9:33249
250
K. Kratena y M. Wüger
Estudios de Economía Aplicada, 2003: 243-258 • Vol. 21-2
distribution used in (15) for energy services is also used for total consumption
expenditure, thus:
C = x
1
C
S
+ (1 - x
1
) C
N
(16)
By these cross section data from the household survey demand shifts are to be
integrated in the analysis. The data is complemented by information from energy
statistics, from the transport databank at the Technical University in Graz, and from
other relevant sources.
Equations (9), (11), (13) and (14) outlined above provide the basis for the econometric
estimates. The results for capital stocks and energy services are shown in Table 1.
Table 1: Capital stock and energy services
Captial stock, K Energy services, SE
Dependent variable Dlog(FA) log(PKM)
Independent variable
log(YD/PC) 0.177 0.094
(0.074) (0.054)
log(PC42) + log(PC44) -0.027
(0.016)
log(FA
-1
) -0.091
(0.073)
log(PKM
-1
) 0.788
(0.052)
log(BEVD2/BEVD1) 0.111
(0.027)
log(PC/PC4) 0.344
(0.049)
Corrected R² 0.577 0.998
Durbin-Watson 1.880 1.896
FA Passenger car stock
PC Price index for private consumption
PKM Passenger km total PC42 Price for fuel
POP Population PC44 Price of the fixed costs for passenger cars
YD Disposible income, current prices BEVD2 Population density outside the metropolitan area
BEVD1 Population density in the metropolitan area
PC4 Price for transport
For the category transport, total demand for transport services is split into private
transport and public transport. These represent the two types of “technologies” which
can be combined to achieve the desired level of mobility demand. The modal split
depends on the relationship between the capital infrastructure networks (road network,
public transport network) and on the relative prices of the two technologies. An additional
calculation was made in order to specify the demand for both forms of technology as
Articulo-04.p65 22/07/03, 9:33250
251MODELLING CONSUMER TRANSPORT DEMAND AND SUSTAINABLE DEVELOPMENT
Estudios de Economía Aplicada, 2003: 243-258 • Vol. 21-2
dependent on relative prices and population density (ignoring the relationship between
the infrastructure capital stocks). This alternative specification revealed that the relative
population density not only has an impact on the total demand for mobility services but
also on the modal split.
After dividing the demand for mobility into these two types of technologies the fuel
consumption for motorised private transport is calculated by multiplying kilometres
travelled by the figure for average fuel consumption of the total vehicle fleet. The
vehicle tax exerts a dampening effect on average fuel consumption. The consumption
category “car purchase” results directly from the changes in capital stock (vehicle
fleet) and the rate of depreciation. Apart from the variable costs of private motorised
transport, as in Conrad and Schröder (1991), we also look at fixed costs, depending on
both capital stock (vehicle fleet) and the price index for fixed costs.
Table 2: Energy demand of private transport
Transport by passenger cars; EN
Passenger km by passenger cars Passenger km by public transport
Excluding Including Excluding Including
population population population population
density density density density
Dependent variable log(PkwKM/PKM) log(PNVKM/PKM)
Independent variable
log(KPkw/KPNV) 0.344
(0.173)
log(PC42/PC43) -0.068 -0.105
(0.031) (0.029)
log(PkwKM
-1
/PKM
-1
) 0.486
(0.138)
log(BEVD2/BEVD1) 0.173 -0.330
(0.044) (0.086)
log(KPNV/KPkw) 0.457
(0.251)
log(PC43/PC42) -0.098 -0.215
(0.045) (0.057)
log(PNVKM
-1
/PKM
-1
) 0.648
(0.089)
Corrected R² 0.874 0.911 0.931 0.911
Durbin-Watson 2.046 2.197 2.144 1.789
PkwKM Passenger km by passenger cars PC42 Price for fuel
PKM Passenger km total PC43 Price for public transport
PNVKM Passenger km by public transport BEVD2 Population density outside the metropolitan area
PC81 Price for electric energy BEVD1 Population density in the metropolitan area
T Time KPkw Road length in km
KPNV Railroad length in km
Articulo-04.p65 22/07/03, 9:33251
252
K. Kratena y M. Wüger
Estudios de Economía Aplicada, 2003: 243-258 • Vol. 21-2
4. SIMULATION RESULTS ON SUSTAINABLE CONSUMPTION
Given the economic consumption model and the econometric estimates as presented
in the previous section, sustainability scenarios for Austria were formulated and “ex
post” simulations were carried out for the period 1990 to 1998. Sustainability scenarios
were defined rather pragmatically: Consumption in the area of transport was said to be
sustainable when over a ten year period it led to a reduction of households’ CO
2
emissions
of 13% compared to the level in 1990 (Austria’s Kyoto target). The results of the ex
post simulations reveal to what extent actual emissions diverge from the target set. The
gap between the two time paths (actual vs. “sustainable” emissions) shows that
substantial intervention would be needed in order to redirect consumption patterns. We
are well aware that the use of such a one dimensional criterion for defining sustainability
cannot capture all relevant aspects related to sustainable consumption. Nevertheless, it
still helps to clarify which structures need to be considered and implemented in economic
modelling of sustainable consumption patterns.
The complex nature of sustainable consumption i.e. the multitude of factors
influencing the formation of consumption patterns makes it necessary to adopt a wide-
ranging bundle of diverse policy instruments. Roughly speaking, available policy
instruments, some of which were used in the simulations described below, can be divided
into three categories (UNEP, 2001; OECD, 2000):
• Economic instruments like tradable emission permits, taxes, direct or indirect
subsidies for specific activities or the removal of environmentally counter-productive
support measures, that exert their influence via price changes.
• Regulatory instruments like typical command and control methods, emission
standards (energy efficiency levels, specific technologies)
• Social instruments – a category of “soft” measures designed to redirect consumer
activity through communication or awareness raising programmes or encouraging
participation in decision making processes, the development of pilot projects etc.
The empirical analysis for Austria shows to what extent changes in specific
consumption areas would have to be made, in order to induce, over approximately a
decade, a move towards more sustainable consumption. Such changes result, e.g. from
policy interventions or from alterations in social values. The model used here has two
main characteristics, where policy interventions and social changes leading towards
sustainability can be implemented:
(i) the explicit modelling of the technology of household production, which can be
changed via price changes or changes in infrastructure (in a broad sense),
(ii)the explicit modelling of different household groups and their consumption patterns
in terms of sustainability.
Technological innovation is likely to contribute as much as demand management
policy. To aid comparison between the scenario results, “maximum settings” were
used in the simulation runs, i.e. in each of the scenarios, and thus for every exogenous
Articulo-04.p65 22/07/03, 9:33252
253MODELLING CONSUMER TRANSPORT DEMAND AND SUSTAINABLE DEVELOPMENT
Estudios de Economía Aplicada, 2003: 243-258 • Vol. 21-2
intervention, the sustainability target had to be reached (the above mentioned 13%
reduction in CO
2
emissions). Furthermore, the required changes in technology, prices
and behaviour are thus made visible. The results indicate that any attempt to achieve
the targeted 13% reduction in CO
2
emissions by concentrating on specific single
measures would involve intense or excessive effort. It seems that moving prevailing
consumption behaviour towards a more sustainable form is only likely to succeed
when a bundle of diverse policy instruments is used.
With respect to transport, the following sustainability scenarios were defined and
calculated:
(1) “Road Pricing”: A kilometre fee for cars is introduced and then returned to
households through a lump sum payment (“eco-bonus”).
(2) “Zero Charge in public transport”: The price of public transport is reduced, the
missing revenues are financed through an increase in vehicle tax.
(3) “Regional Planning”: This scenario assumes an increase in the density of city
population.
(4) “Demand shifts”: The share of “normal” households, measured in terms of the
structure and amount of consumer expenditure, decreases in favour of “sustainable”
households.
Sustainability Scenario “Road Pricing”
This scenario comes closest to a neoclassical environmental policy approach as it
targets emission prices. A tax on kilometres travelled is introduced. This tax could be
implemented in the form of a vehicle charge per kilometre. The crucial aspect in this
scenario is revenue neutrality (i.e. tax revenues raised are returned lump sum to
households). The recycling of the revenues should prevent any negative macro-economic
impact. Ex post, i.e. after the desired change in demand, the road pricing policy generates
1.8 bn for the approximately 45 bn km driven (average 1990 – 1998). This represents
an effective charge of about 0.04 per kilometre.
The road charge raises consumer prices by about 1.7%, while the return of the
revenue to households has an opposite effect so that the net impact is close to zero. As
a result, there is hardly any change in real household disposable income. Thus, given
the fall in real expenditure on individual motorised transport, we see a corresponding
increase in expenditure on non-energy consumption. Further, not only has fuel
consumption fallen by 11.3% necessary to achieve the targeted CO
2
levels, all forms
of expenditure on individual motorised transport fall too. This includes car purchases
(approx. -15%) and vehicle fixed costs (-7.5%) since, similar to an increase in fuel tax,
the road charge raises the variable costs of car use. Since the decline in the demand for
public transport is only 0.9%, we also see a change in the demand structure for transport,
with an alteration in the modal split between private and public transport. Summing
up, the “road pricing” scenario, in correspondence with the many other studies on
energy taxation, shows that with respect to the standard indicators employed in national
Articulo-04.p65 22/07/03, 9:33253
254
K. Kratena y M. Wüger
Estudios de Economía Aplicada, 2003: 243-258 • Vol. 21-2
1. In 1993 an insurance tax based on engine size has been introduced.
accounts, sustainable development could indeed be combined with positive macro-
economic effects (double dividend).
Sustainability Scenario “Zero Charge in Public Transport”
This scenario is based on the assumption that the general acceptance of public
transport will increase. The extent to which cost considerations affect such acceptance
is also tested here. Thus, this scenario is also an example of neoclassical price regulation
of emissions being used as an environmental policy instrument. The specific design of
the scenario does, however, allow additional aspects to be captured, but these are only
dealt with qualitatively and not quantitatively. In this sustainability scenario, the price
for public transport is reduced on average over the period 1990 to 1998 by 30%, and
the resulting revenue loss experienced by the transport companies is offset by an increase
in vehicle tax. This can be seen as the provision of a cross-subsidy for public transport
financed by the vehicle tax. As a small proportion of vehicle tax
1
is earmarked for
developing the public transport network, cross-subsidisation of public transport
operations would imply only a rather minor change in the system.
The required fall in fuel consumption of 11.3% (given the CO
2
target) is achieved
through the shift from private to public transport (+6.1%), and also through the various
effects of the increase in vehicle tax on fuel consumption. The measure, which (ex
post) generates extra tax income of 363 million, has an impact on the average fuel
consumption of the car fleet, and influences the real fixed costs of private transport,
which in turn dampens vehicle purchases. I.e. fewer, and at the same time, more fuel-
efficient vehicles are purchased in an effort to offset, at least partly, the increase in
vehicle tax. The combined price and income effects result in a real decline of car
purchases by 15% and a real decline in fixed costs related to car ownership by 25%. In
total, we see here a smaller increase in non-energy consumption (+1.3%) than in the
“road pricing” scenario. Thus, in this scenario too, positive effects on non-energy
consumption and accompanying advantages for the macro-economy are to be expected.
This scenario differs from “road pricing” with respect to the compensatory payments
(lump sum transfer) to offset the increased tax and to achieve a double dividend effect.
Here, a double steering effect is achieved by targeting compensation specifically at a
lowering of public transport prices.
In order to raise the attractiveness of public transport and to increase the steering
effect still further, additional measures should be used at the same time (e.g. schemes
to promote co-operation with taxi operators). It is also possible to make the idea of
cost redistribution even more transparent for households by providing a direct link
between the two cost elements. For example, the annual payment of vehicle tax could
be automatically coupled with the provision of a public transport pass.
Articulo-04.p65 22/07/03, 9:33254
255MODELLING CONSUMER TRANSPORT DEMAND AND SUSTAINABLE DEVELOPMENT
Estudios de Economía Aplicada, 2003: 243-258 • Vol. 21-2
Sustainability Scenario “Regional Planning”
This scenario is based on the changes in lifestyles that occurred between 1990 and
1998 and which led to an increase in traffic volume. The central variable in the model
is the ratio of the population density in rural areas (surrounding cities) to that in the
cities. This ratio rose continuously throughout the period 1990–1998, reflecting a
lifestyle of “work in the city - live in the country”. In contrast to this consumer lifestyle,
we assume that population movement would have led to increased concentration of
residential and work areas within cities. This leads to a significant rise in city population
density, with the respective figures going up by 40 persons per square kilometre (+29%).
This sustainability scenario highlights the interface between transport and housing.
Regional shifts in the population have no impact on the total level of private consumption.
This scenario reflects the idea of reducing redundant energy services without a decrease
in the level of economic welfare. Since, however, only expenditure on fuel consumption
falls and other transport related expenses remain almost constant, we witness merely a
slight shift between energy and non-energy consumption.
Sustainability Scenario “Demand Shifts” in Transport
This scenario analyses the effects of an increased share of “sustainable” households.
The distinguishing characteristics are changes in transport behaviour. In order to achieve
the CO
2
target in this scenario, the two household types – as described above - are no
longer grouped by the median value. An emission reduction of 13% would call for a
share of 64% of households to consume in a more sustainable way, and only 36%
could maintain their conventional consumption behaviour. This assumption leads to
an expenditure shift between private and public transport with aggregate transport
costs remaining constant. In total, this leads to a slight increase (+3%) in kilometres
travelled. The analysis of consumption expenditure for each household type showed
that “sustainable” households also tend to have lower total consumption expenditure
than “normal” households (given the same income). The shift in the household structure
thus leads to a decline in total real private consumption by about 2%, whereby non-
energy consumption falls by 2.6% (see Table 3).
5. CONCLUSIONS
The focal point of the analysis is consumer services derived from a combination of
stocks (e.g. building stock, transport systems) and flows (mainly energy). This approach
intends to illustrate that a concentration on flows in consumption models lacks important
aspects of sustainable consumption.
Articulo-04.p65 22/07/03, 9:33255
256
K. Kratena y M. Wüger
Estudios de Economía Aplicada, 2003: 243-258 • Vol. 21-2
The economic analysis of sustainable consumption patterns aims to clarify the
extent to which it might be both possible or commendable to promote the substitution
of flows by stocks (e.g. more energy efficient transport systems). Relevant too, in this
respect, is the role that can be played by technological innovation (improvements in the
existing vehicle fleet stock, or the potential for introducing specific incentives to promote
new technologies). Two essential factors are crucial in the context of sustainable
consumption: the demand shifts concerning the consumer services desired, and the
composition of the stock-flow mix necessary for the service provision.
The empirical evidence produced by the model reveals that economic instruments –
in this case taxes and road pricing – can generate positive macro-economic effects.
This result is consistent with the evidence of other studies that have been undertaken to
test the effects of incentive based instruments in environmental policy.
The model framework also allows to assess the economic impact of shifts in demand.
In the concrete case, as far as the demand for energy in transport is concerned, the
model is capable of determining how high the shift in the share of Austrian households
towards more sustainable consumption would have to be, in order to reach the target
set. The explanatory and predictive power of the model are largely exhausted in
answering how these shifts in demand are propagated or how large monetary incentives
need to be in order to induce such changes in consumption behaviour (demand shifts).
Table 3: Simulation Results for the Sustainability Scenarios “Transport”
(Average 1990 - 1998)
Road Zero Regional Demand
pricing charge planning shift
Difference to baseline in %
Consumer prices 1.7 1.3 - -
Private consumption. total 0.1 -1.1 0.0 -1.9
Non-energy consumption 2.0 1.3 0.5 -2.6
Private consumption (P 95)
Transport
Cars -14.6 -15.0 0.0 -
Fuels -11.3 -11.3 -11.3 -11.4
Public transport -0.9 6.1 -0.4 20.0
Other transport -7.5 -24.7 0.0 -
Transport activities
Person-kilometres. total -12.7 -10.4 -12.5 2.8
Person-kilometres. cars -17.9 -15.6 -17.9 -17.9
Person-kilometres. public
transport -1.7 0.7 -0.6 45.8
Articulo-04.p65 22/07/03, 9:33256
257MODELLING CONSUMER TRANSPORT DEMAND AND SUSTAINABLE DEVELOPMENT
Estudios de Economía Aplicada, 2003: 243-258 • Vol. 21-2
Looking at the issues in qualitative rather than quantitative terms, however, it can be
assumed that apart from changes in consumption, the realisation of structural change
on the supply side should also prove possible. The interactions between technology
and changes in public awareness becomes clear. Such complexity makes it all the more
necessary that political objectives are stated as explicitly as possible.
Gratitude
This article resulted from the study “Kletzan, D., Köppl, A., Kratena, K., Wüger, M., Economic
Modelling of Sustainable Structures in Private Consumption and An Analysis of Heating and Transport,
Study by the Austrian Institute of Economic Research commissioned by the Federal Ministries for
Agriculture, Forestry, Environment and Water Management and for Transport, Innovation and
Technology, WIFO, Vienna, 2002”.
REFERENCES
BECKER, G. S., “A theory of the allocation of time”, Economic Journal, 1965, 75, S 493-517.
BOSSEL, H., “Policy assessment and simulation of actor orientation for sustainable development”,
Ecological Economics, 2000, 35 (3), 337 – 355.
BROWN, P. M., CAMERON, L. D., “Survey: What can be done to reduce overconsumption?”
Ecological Economics, 2000, 32 (1), 27 – 41.
C
ONRAD, K., SCHRÖDER, M., “Demand for durable and nondurable goods, environmental policy
and consumer welfare”, Journal of Applied Econometrics, 6, 1991.
DEATON, A, MUELLBAUER, J., Economics and consumer behavior, Cambridge, 1980.
DUCHIN, F., Structural economics: measuring change in technology, lifestyles, and the
environment, Island Press, 1998.
GINTIS, H., “Beyond Homo economicus: evidence from experimental economics”, Ecological
Economics, 2000, 35 (3), 311 – 322.
GRAHAM, J.W., GRUN, C.A., “Estimating the Paramesus of a house hold production fundion
roidh Jaird Products”, the Rivino of Economics and Stadistics, 1984, p. 277-282.
JAGER, W., JANSSEN, M. A., DE VRIES, H. J. M., DE GREEF, J., VLEK, C. A. J., “Behaviour in
commons dilemmas: Homo economicus and Homo psychologicus in an ecological-
economic model”, Ecological Economics, 2000, 35 (3), 357 - 379.
JAGER, W., JANSSEN, M. A., VLEK, C., “Experimentation with household dynamics: the consumat
approach.”, Sustainable Development 2001, (1), 90-100.
KLETZAN, D., KÖPPL, A., KRATENA, K., WÜGER, M., Economic Modelling of Sustainable
Structures in Private Consumption – An Analysis of Heating and Transport, Study by the
Austrian Institute of Economic Research commissioned by the Federal Ministries for
Agriculture, Forestry, Environment and Water Management and for Transport, Innovation
and Technology, WIFO, Vienna, 2002.
LANCASTER, K.J., “A new approach to consumer theory”, Journal of Political Economy, 1966,
74, S 132-57.
Articulo-04.p65 22/07/03, 9:33257
258
K. Kratena y M. Wüger
Estudios de Economía Aplicada, 2003: 243-258 • Vol. 21-2
OECD, Towards more Sustainable Consumption: An Economic Conceptual Framework, Working
Party on Economic and Environmental Policy Integration, Paris, 14-15 November 2000
RØPKE, I., “Analysis: The dynamics of willingness to consume”, Ecological Economics, 1999,
28 (3), 399 – 420.
RØPKE, I., “Is consumption becoming less material? The case of services.”, Sustainable
Development, 2001, (1), 33-47.
ROTH, T.P., The Present State of Consumer Theory. The Implications for Social Welfare Theory,
Lanham, 1998, University Press of America.
SHOVE, E., WARDE, A., Noticing inconspicuous consumption, Paper presented at the ESF
TERM Programme workshop on Consumption, Everyday Life and Sustainability, Lancaster
University, April 1997.
SIEBENHÜNER, B., “Commentary: Homo sustinens – towards a new conception of humans for
the science of sustainability”, Ecological Economics, 2000, 32(1), 15 – 25.
STIGLER, G. J., BECKER, G. S., “De Gustibus Non Est Disputandum”, The American Economic
Review, 1977, 67(2), 76-90.
UNEP, Consumption Opportunities, Strategies for change, A report for decision-makers, Genua,
2001.
WENKE, M., “Umweltbewusstsein und Konsumverhalten der privaten Haushalte - Theorie und
Evidenz am Beispiel der Nachfrage nach Haushaltschemikalien”, RWI Mitteilungen, 44(1),
1993.
WILHITE, H., NAKAGAMI, H., MASUDA, T., YAMAGA, Y., HANEDA, H., “A cross-cultural analysis
of household energy-use behavior in Japan and Norway”, Energy Policy 1996, 24(9),
795-803.
WILK, R., Culture and Energy Consumption, Indiana University, Bloomington, mimeo, 2001.
WILLET, K. D, NAGHSHPOUR, S., “Residential demand for energy commodities: A Household
Production Function Approach”, Energy Economics, Oktober 1987, 9(4), 251-56.
Articulo-04.p65 22/07/03, 9:33258
