Integrated passenger transport policy assessment within a computable general equilibrium framework [original]

Integrated passenger transport policy assessment within a computable general

equilibrium framework

vorgelegt von

Diplom-Volkswirtin

Dominika Kalinowska

geboren am 9. Mai1976 in Lodz, Polen

Von der Fakultät VII – Wirtschaft und Management

der Technischen Universität Berlin

zur Erlangung des akademischen Grades

Doktor der Wirtschaftswissenschaften

Dr. rer. oec.

genehmigte Dissertation

Promotionsausschuss:

Vorsitzender Prof. Dr. Christian von Hirschhausen

Berichter: Prof. Dr. Georg Meran

Berichter: Prof. Dr. Karl W. Steininger

Tag der wissenschaftlichen Aussprache: 02. März 2010

Berlin 2010

D 83

Integrated passenger transport policy assessment within a

computable general equilibrium framework

vorgelegt von Diplom-Volkswirtin Dominika Kalinowska

an der Fakultät Wirtschaft und Management der Technischen Universität Berlin

zur Erlangung des akademischen Grades Dr. rer. oec.

Berichter: Prof. Dr. Georg Meran (Technische Universität Berlin)

Berichter: Prof. Dr. Karl W. Steininger (Universität Graz)

Source: http://www.torontorealtyblog.com/wp-content/uploads/2009/12/tollbooth.jpg.

Berlin 2010

Zusammenfassung

Gegenstand der vorliegenden Dissertation ist die Entwicklung eines methodischen

Werkzeuges, konkret eines allgemeinen rechenbaren Gleichgewichtmodells (Computable

General Equilibrium Model, CGE), welches zur Bewertung politischer Maßnahmen den

Personenverkehr bzw. die Personenverkehrsleistung in Deutschland im

volkswirtschaftlichen Kontext abbildet. Ziel der Arbeit ist die Anwendung dieses

methodischen Instrumentariums zur Evaluierung politischer Maßnahmen im

Verkehrssektor – wie der Einführung von Straßenbenutzungsgebühren für Pkw –

hinsichtlich resultierender Effekte, einschließlich der gesamtwirtschaftlichen Folgen.

Zur Untersuchung ökonomischer Umverteilungseffekte einer Pkw-Maut werden im Modell

nach Einkommens- und Wohnsiedlungsstruktur differenzierte Haushaltskategorien

abgebildet. Dabei werden erhebungsbasierte Mikrodaten zur Mobilität und

Einkommensverwendung privater Haushalte in ein rechenbares, allgemeines

Gleichgewichtsmodell der gesamten deutschen Volkswirtschaft integriert. Damit wird die

private Nachfrage nach motorisiertem sowie öffentlichem Verkehr in dem CGE Modell

über realisierte Konsum- bzw. Nachfrageentscheidungen privater Haushaltskategorien

abgebildet. Als wichtiges Ergebnis zeigt die Arbeit regressive Wirkung einer Pkw-Maut

innerhalb der Einkommensverteilung privater Haushalte in Deutschland. Gleichzeitig

würde eine Pkw-Maut von 5 Cent/km die CO

Emissionen in diesem Sektor bereits um bis

zu 10 % senken. Die regressive Wirkung der Maßnahme kann durch eine adäquate

Ausgestaltung der Einnahmenverwendung kompensiert werden. Damit kann die oft als

fehlend diskutierte Akzeptanz der Pkw-Maut innerhalb der Bevölkerung erhöht werden.

Schlagwörter:

Personenverkehr, Allgemeine Gleichgewichtsmodellierung, Wohlfahrtseffekte,

Umverteilungspolitik, Gerechtigkeit, Straßennutzungsgebühren

Abstract

The purpose of this dissertation work is the development of a methodological framework in

terms of a computable general equilibrium (CGE) model in order to carry out an integrated

impact assessment from road charging introduced in the passenger car sector for Germany.

The pricing policy measure is applied to the private motorized travel demand covering the

overall road network. The uniqueness of this work distinguishing it from the current state of

the research in the area of CGE modeling is firstly the application of an existing

methodological approach to a newly constructed database for Germany, secondly the

extension of the model framework by integrating land use characteristics of private

household residential location, and finally the funded assessment of distributional and

equity implications within the private household sector from the introduction of road use

charging. To give a better understanding of distributional, equity and welfare impacts from

the introduction of car road pricing within the overall economic context, mode specific

travel demand of private households is integrated into the CGE model. The modeling

framework accounts for different household categories with respect to income and

residential location through the integration of behavioral mobility parameters as well as

household travel expenditures. The analyses of policy simulations are carried out

introducing different road charging revenue redistribution schemes. The results of this work

show that distributional effects and equity concerns are strongly related to the revenue use

patterns as well as to country and household specific travel demand profiles. However, the

introduction of road user charges can have a positive impact on environmental welfare

through the reduction of car use and the corresponding CO

emissions, when mode shifts

are induced through the implementation of the policy reform. They results of this work

provide country specific insights concerning the public acceptance of road user charges

applied to passenger car travel.

Keywords:

Passenger road travel, Computable general equilibrium model, Transport policy

assessment, Road use charging, Distributional impacts, Equity

Table of Contents

1 Introduction....................................................................................................................19

2 Theoretical discussion....................................................................................................23

2.1 Theoretical discussion of road pricing....................................................................23

2.1.1 Social, private, internalised, and external costs of road use.............................23

2.1.2 Marginal social cost pricing principles.............................................................26

2.1.3 Alternative approaches to marginal social cost pricing....................................33

2.1.4 Policy frameworks and road pricing calculations.............................................36

2.1.5 Introduction of road charging in Germany.......................................................39

2.2 Discussion: acceptability of road charging.............................................................42

2.3 Equity concepts and criterions ................................................................................45

3 Computable general equilibrium (CGE) and travel demand modelling – extension of

the research...........................................................................................................................55

4 Model and database .......................................................................................................56

4.1 The general computable equilibrium model............................................................56

4.2 CGE and travel demand modelling.........................................................................59

4.2.1 Description of the German Road Travel Policy Model (GRTPM) ..................61

4.2.2 Private household sector, travel demand and transport production..................65

4.3 Database construction .............................................................................................70

4.3.1 Input-output data ..............................................................................................70

4.3.2 Transportation sector input-output data............................................................77

4.3.3 Mobility data ....................................................................................................80

4.3.4 Household budget and expenditure data...........................................................81

4.3.5 Combining different data sources – private household income, expenditures,

and mobility...................................................................................................................86

4.3.5.1 Construction of equivalence-weighted income quartiles...........................87

4.3.5.2 From categorical to continuous income data in the National Travel

Survey...........................................................................................................90

4.3.5.3 Matching different data sources.................................................................92

5 Model implementation...................................................................................................96

5.1 Policy scenario definition........................................................................................96

5.2 Revenue reallocation schemes – household refund.................................................99

5.3 Model results.........................................................................................................101

5.3.1 Overall transport and macroeconomic impacts..............................................101

5.3.2 Environment...................................................................................................105

5.3.3 Sectors ............................................................................................................105

5.3.4 Budget.............................................................................................................106

5.3.5 Microeconomic impacts .................................................................................106

5.3.5.1 Household travel expenditure..................................................................107

5.3.5.2 Household travel demand ........................................................................122

5.3.5.3 Households’ contribution to the reduction in CO

emissions..................135

5.3.5.4 Household distributional welfare and equity effects ...............................138

5.3.5.5 Equity measurement ................................................................................142

6 Conclusion...................................................................................................................145

7 Bibliography ................................................................................................................153

8 Appendix......................................................................................................................175

List of Figures

Figure 1 Optimal taxation in the presence of social costs from car road travel with a

basic road pricing model.................................................................................27

Figure 2 Structure of household h final demand ...........................................................68

Figure 3 Symmetric input-output table at basic prices (product by product)................72

Figure 4 Social accounting matrix of the GRTPM – attribution of private household

travel expenditures to the sectoral input-output database ...............................78

Figure 5 Contents of the Mobility in Germany 2002 survey.........................................81

Figure 6 Basic regional settlement structure typification” (Siedlungsstrukturelle

Regionsgrundtypen), German Federal Office for Building and Regional

Planning (BBSR).............................................................................................85

List of Tables

Table 1 Overview of social and private cost components of car use ...........................25

Table 2 Sector data of the GRTPM..............................................................................76

Table 3 Information available on household income comparing the Household

Income and Expenditure Survey (EVS, 2003) and the German National

Travel Survey (MiD, 2002).............................................................................88

Table 4 Equivalence-weighted household income quartile intercepts calculated from

the Sample Survey of Income and Expenditure and the National Travel

Survey..............................................................................................................92

Table 5 Household categories used in the application of the GRTPM........................93

Table 6 Selected statistics from the Survey of Income and Expenditure and the

National Travel Survey...................................................................................94

Table 7 Overview of the policy scenarios implemented in the GRTPM ...................100

Table 8 Macroeconomic effects from different road charging scenarios, Germany

2002...............................................................................................................102

Table 9 Car and public transport expenditures across household categories.............107

Table 10 Car and public transport expenditure impacts across household categories

and road charging revenue reallocation scenarios.........................................108

Table 11 Household income, overall consumption and selected transportation

expenditures for household categories..........................................................111

Table 12 Household overall consumption and selected transportation expenditures

as shares in household net income for household categories........................113

Table 13 Household expenditures on road charging and fuel tax as share in income

for different household categories.................................................................116

Table 14 Level of road charging payment and household refund according to

scenario B in million Euro for different household categories .....................117

Table 15 Household share of the road charging payment and the policy refund in

the revenue and refund total according to scenario B in % for different

household categories.....................................................................................119

Table 16 Road charging refund redistributed to private households as share in

income, different household categories.........................................................120

Table 17 Road charging refund redistributed to private households as share in the

amount of the road charging payment, different household categories ........122

Table 18 Car, public transportation, and total distance travelled in km per year and

per household across household categories...................................................123

Table 19 Distribution of car, public transportation, and total distance travelled

across household categories..........................................................................124

Table 20 Overall kilometres travelled including all modes per day and per

household by household type and trip purpose.............................................126

Table 21 Household distribution as to residential location and the presence of

children in the household ..............................................................................127

Table 22 Share in % of single households in the total number of households

according to household category...................................................................128

Table 23 Car and public transportation distance travelled in km per household and

per workday for household categories ..........................................................129

Table 24 Household shares according to motorization level for different household

categories.......................................................................................................130

Table 25 Car and public transportation distance travelled impacts across household

categories and road charging revenue reallocation scenarios .......................132

Table 26 Selected CO

emission characteristics of household categories in the pre-

policy situation, Germany 2003....................................................................136

Table 27 Overall CO

emission impacts across household income groups and road

charging revenue reallocation schemes.........................................................137

Table 28 Welfare impacts across household categories and road charging revenue

reallocation scenarios....................................................................................139

Appendix

Appendix 1 List of core model equations.......................................................................175

Appendix 2 Variables.....................................................................................................177

Appendix 3 Gross domestic product calculation approaches for an input-output table.180

Appendix 4 Input-output tables and data sources within the national and European

system of accounts......................................................................................181

Appendix 5 Household refund distribution in scenario B according to fuel tax

payments.....................................................................................................182

Appendix 6 Household expenditures on fixed car use related components for

household categories...................................................................................183

Appendix 7 Household expenditures on transportation as shares in the overall

consumption expenditure for household categories....................................184

Appendix 8 Household expenditures on fixed car use related components as shares

in household net income for household categories.....................................185

Appendix 9 Household expenditures on fixed car use related components as shares

in household overall consumption expenditure for household categories..186

Appendix 10 Accessibility of centres and functional urban areas in Germany by car,

German Federal Office for Building and Regional Planning (BBSR) .......187

Appendix 11 GAMS code of the German Road Travel Policy Model (GRTPM), 2002 .189

Vorwort

Die Idee zu dieser wissenschaftlichen Arbeit entstand hauptsächlich entlang der

Herausforderung, Inhalte und Empirie aus der angewandten ökonomischen Verkehrs-

forschung der Abteilung Energie, Verkehr, Umwelt des DIW Berlin mit einer innovativen

Methodik von angewandten, rechenbaren allgemeinen Gleichgewichtsmodellen unter dem

„Mantel“ einer politisch und gesellschaftlich (aktuell) relevanten Fragestellung zu vereinen.

Die Aufbereitung der Datenbasis erwies sich als zeitaufwendig und nicht frei von Mühen –

unterschiedliche Datenquellen mussten hierfür herangezogen, ineinander integriert und zum

Teil mithilfe ökonometrischer Verfahren ergänzt werden. Die Ausweitung der konkreten

Fragestellung nach Wohlfahrts- und Umverteilungseffekten aus politischer

Maßnahmenimplementierung um die normative Komponente nach ihrer Fairness erhöhte

die Komplexität der Arbeit. Möge der Leser beurteilen, wie gut die aufgeworfenen Fragen

in der Ausarbeitung beantwortet wurden.

Jedenfalls wäre diese Arbeit nicht ohne die Unterstützung einer ganzen Reihe von Personen

möglich gewesen, bei denen ich mich auch noch mal an dieser Stelle herzlich bedanken

möchte. Professor Karl Steininger hat das Ausgangsmodell für Österreich zur Verfügung

gestellt und bei meinen vielen Fragen Hilfestellung gegeben. Besonderer Dank gilt

Professor Georg Meran für seine Unterstützung dieses Forschungsprojektes. Ebenso hat mir

meine Abteilungsleiterin Professor Claudia Kemfert ganz erheblich bei der Durchführung

des Promotionsvorhabens geholfen. Mein ganz besonderer Dank gilt auch meinem lieben

Kollegen Hartmut Kuhfeld.

Dominika Kalinowska

Dezember 2009

1 Introduction

Policies of any kind – targeted or not targeted – influence demand. Growing concerns about

energy consumption and supply issues, environmental externalities, and climate change

matters emerge in the need for policy involvement. These concerns are closely linked to

transportation, in particular the road travel sector. Hence, policy instruments are needed.

They are being developed and partially implemented to tackle negative external impacts

from motorised road travel on the environment and the society. Their effects require an

assessment and a deeper understanding. The objective to provide a deeper understanding of

the overall economic and environmental effects from car road charging in general and

distributional impacts in particular are the driving motivation for this dissertation work. The

development and the application of the methodological tool for an integrated assessment of

such effects – with strong focus on household distributional and equity effects – is the core

research task of this dissertation work.

There are several motivations for constructing a model for Germany: Obviously Germany

is a large and therefore important economy. It also offers a good coverage of data that are

required for the construction of the model database, i.e., input- output- and social

accounting data as well as household travel, income and expenditure data. Germany is

furthermore one of the countries, where the idea or the proposal for the introduction of car

road charging shows up regularly on the policy agenda. Arguments propagating car road

pricing are to finally introduce a road use charge on foreign drivers using German road

network otherwise without paying for it. Furthermore, because a distance dependant road

use charge applied to the freight sector is already in place since 2005, it is not unlikely to

assume that with manageable technological enhancements the existing revenue collecting

infrastructure in the freight transportation sector could be adapted to passenger cars. The

technical feasibility of the measure is a strong criterion when discussing its realisation; an

introduction of car road charging as discussed later on in this work is not an unrealistic

undertaking.

Taxes or charges levied on car users are often designed to meet specific internalisation

objectives of uncovered (external) costs associated with car use. In some cases they

combine the price of negative environmental or social impacts from (fossil) fuel

combustion and road use together with the payment for road infrastructure provision. No

matter what is their exact definition, road charging measures share a common problem:

they lack (broad) public acceptance. Introduction of a road use charges or an augmentation

of the fossil fuel tax evokes strong welfare distributional and equity concerns, mainly

among car users. Broad public opposition to car use related measures for reduction of fossil

fuel consumption and therefore climate change mitigation is a well-known phenomenon

and often discussed in the literature. However, the effective scope of the impacts on private

car users and the overall economy from the augmentation of automobility cost depends on

the exact design of the measure and on the accompanying revenue use or reallocation

scheme. Relevant factors are therefore the magnitude and range of the policy measure or

the level and implementation scheme of the charge – on what assumptions is it based – and

what are the integrated (welfare) impacts on private households, as well as effects on

environment and the overall economy. Still, it is common presumption and concern that

distance dependent road use charges work regressively regarding their economic and social

impacts. As households from the low income categories spend a greater proportion of their

income on travel and energy, they are more likely to be adversely affected by imposed

charges on car road use. In practice, the actual impact of road user charging on each

economic sector or agent is closely linked to specific characteristics of the national

economy, but also to car travel demand patterns conditioned upon the observable and

implicit sociodemogarphic and economic structure of private households, accessibility and

land use attributes. The strong influence of the initial living conditions of households on

policy outcomes suggests the application of an assessment instrument that would account

for these characteristics through the model specification and the construction of the

database.

The methodological approach has to take into account the multidimensional and integrated

impact of policy measures such as road use charging, especially when public revenue is

collected and reallocated. The objective of the presented doctoral thesis is therefore to

investigate equity and welfare distributional effects from car road charging on the private

household sector in the specific case of Germany. The policy impact assessment covers the

entire economy with special focus on the private households and their welfare situation

before and after the introduction of the measure. To give a funded understanding of the

mode of functioning of environmental and transport policy measures targeting car use,

especially on their socioeconomic implications, an overall model of the German economy

is extended by a more disaggregated private household representation. The methodology

used in this work is a computable general equilibrium framework with integrated private

household car travel and public transport demand, differentiated by combined income and

spatial residence characteristics.

The work reveals the effectiveness of road use charging being the only approach of its kind

carried out for Germany. Varying scenario assumptions as to the design of the policy

reform provide the basis for the simulation analysis of the model. As a result the work

quantifies the relationship between welfare and equity effects from road user charging and

the formulation of the revenue reallocation schemes. The application of household category

specific behavioural parameters in form of travel demand and travel demand elasticities

approximate the practical real life situation of the policy implementation environment (in

Germany). Results obtained are derived from a sound empirical basis providing a new

insight in the integrated assessment modelling research. The uniqueness of this work

distinguishing it from the current state of the research in the area of CGE modeling is

therefore, firstly the application of an existing methodological approach to a newly

constructed database for Germany, secondly the extension of the model framework by

integrating household categories of equivalent income quartiles and land use characteristics

of households’ residential location, and finally the quantitative assessment of distributional

and equity implications within the private household sector from the introduction of road

use charging.

In Chapter 2 of this thesis work relevant theoretical aspects for the methodological

application and the interpretation of the results are discussed. Research work on the basic

economic concept behind road charging, on road charging acceptability, and on welfare,

distributive and equity effects within the economic context are presented. Chapter 3

proceeds with an overview of the research done within the area of applied overall economic

impact assessment of road pricing measures in the transport sector, taking into account

welfare distribution and equity effects. The objective underlying this doctoral work and the

description of the methodology used to obtain the results conclude Chapter 3.

Chapter 4 focuses on the detailed description of the model and the data applied. It starts

with a brief introduction of some basics of the general equilibrium theory, goes over the

model structure and its particularities, and ends with the description of the extensive

database and the multiple data sources used to construct the CGE model with integrated

private household demand. Emphasis is put on the specific incorporation of passenger

travel demand into the economic modelling frame.

After the initial definition of varying scenarios underlying the policy simulations, Chapter 5

deals with the description, interpretation and discussion of the model results.

Chapter 6 concludes and gives a critical outlook on the research subject of this thesis work.

2 Theoretical discussion

In the following, relevant theoretical aspects for the methodological approach underlying

this work are discussed. The discussion includes in particular relevant aspects for the

interpretation of the modelling results and for answering the research question concerning

distributional and equity impacts from road use charging in the passenger car sector.

2.1 Theoretical discussion of road pricing

In this Chapter, firstly relevant basics concerning the theoretical background of road pricing

are revealed. The presented theoretical discussion about social cost pricing focuses

exclusively on the road travel sector as the core subject of this work. Some of the

theoretical aspects discussed are then linked to examples of the implementation of the

measure referring to current European Union (EU) policy framework as well as national

policies regarding road pricing. The Chapter concludes linking the theoretical discussion

with the road charging policy implemented later on in this work.

2.1.1 Social, private, internalised, and external costs of road use

Road pricing or road charging defined as direct collection of user charges per km of

infrastructure demand look back at nearly a century of economic research and literature. It

is based on the simple reasoning of economic rationality that the individual willingness to

pay for consumption of goods and services should reflect the marginal social cost of its

production. Market pricing of transportation resources such as vehicles, fuel, insurance, etc.

does not automatically bring about an efficient use of public roads. Direct cost of car

ownership and use represent individual transportation costs and their payment does not

cover the overall social cost caused by (car) transport. Social cost of car travel has to reflect

the monetarised damage or negative externalities imposed through individual car use on

society as a whole as well as on other car and road users. Pigou (1920) based his concept of

externalities on the work of Marshall (1890) and referring to it as the “indirect effect of a

consumption activity or a production activity on the consumption set of a consumer, the

utility function of a consumer or the production function of a producer.”

External costs are

therefore induced by transport users but not borne by them, instead these costs are passed

on to third parties and the general public. The idea of external costs and the implementation

of road pricing goes beyond the mere application of an economic, market based pricing

mechanism to the commodity “road space” reflecting its scarcity. Its application is

extended to such goods as environmental quality, health, safety, even quality of life, and

others. Jansson (1997) defines marginal social cost of infrastructure use as the sum of the

following marginal costs:

Costs borne directly by the motorist (provision of vehicle and fuel, travel time of the car

user),

Cost imposed on the infrastructure provider (provision and maintenance of the

infrastructure),

Costs imposed on the infrastructure users (delays and increased risk of accidents),

Costs imposed on society as a whole (environmental pollution, noise, global warming,

etc.).

Hence, firstly a distinction between private and other social costs is made, where private

costs are most often borne by the road or vehicle user and therefore directly internalised.

The social costs comprise the extern social costs not born the private road user and the

internalised social costs borne by the private road user. The extern cost components of the

social cost of road use are generally used as the basis for the social marginal cost pricing

calculation.

Table 1 summarises the distinction by presenting common negative externalities associated

with motorised travel and (marginal extern) social cost calculations as well as the

individual or private costs of car use born by each user.

External economies are mentioned in Marshall (1890, p. 266).

Table 1 Overview of social and private cost components of car use

Negative externalities their social costs components vs. private costs generated from motorised travel

Negative externalities and extern social cost components

from car use

Private, internalized costs from car use

Congestion cost:

- time losses, i.e., extra time costs caused to other

road users,

- vehicle operating costs,

- environmental pollution, etc.

Cost of car trip before taxes:

- car ownership,

- maintenance,

- fuel, parking,

- insurance, etc.

Health risks, fatalities and infrastructure damage from road

accidents

Own time costs

Air pollution, e.g., CO

emissions Taxes related to car and fuel expenditures

Noise exposure Other car use related charges

Climate change Own accident costs

Other externalities:

- damage of nature and landscape,

- land use,

- soil and water pollution,

- up- and downstream processes,

- additional costs in urban areas,

- energy or fossil fuel dependency of the motor

vehicle sector, etc.

Insurance premium for accident costs caused to others

Sources: Delucchi (1997), Quinet (2004).

The total amount of the social external costs results from the summation of the cost

components enumerated in the first column of Table 1.

Externalities can be basically defined as a situation where the economic activity of one

agent alters the profit or utility of at least one other agent in the economy and is not

accounted for by the market. According to economic principles of the neoclassical

economic theory this leads to a market failure providing a suboptimal or Pareto inferior

solution. Some scholars argue that the growth of road traffic and resulting externalities are

the result of distorted pricing policies for the different modes of transport, in particular road

travel. The failure to account for external costs in transport prices have led to excessive

growth of the more polluting modes of transport and on the other hand hampered the

growth of more environmentally friendly modes.

Hence, many scholars, among them Marshall (1890), Pigou (1920), Smith (1937), Dupuit

(1952 [1844] and 1962 [1849]), Vickery (1948 and 1968), Coase (1946), or Mohring (1964

and 1976), have been arguing in favour of (mainly) government intervention. The existence

of externalities justifies therefore government intervention, e.g., through the introduction of

prices to internalise given externalities.

2.1.2 Marginal social cost pricing principles

The efforts behind the assessment of social road use costs are induced by the need to set

(road use) prices as a method of resource allocation. The argument that the “right” price

does not exist has important implications for the discussion of road pricing where different

cost calculation approaches have been developed over the last century. Hence, pricing is

based on the principle that it should rather reflect optimal pricing strategies that allow

reaching specific goals. The optimal profit maximising price may not coincide with the

price that would lead to the welfare maximum (consumer oriented) vs. one that would

cover long run investment decisions (supplier oriented). The main difficulty of pricing

policies – in particular also in the case of road pricing – lies in finding the general

consensus on the objective or the goal to be achieved by a price setting (Button, 1993a).

The theoretical argumentation behind the market failure due to the uncovered social cost in

case of car and road use is often discussed based on the (graphical) reasoning behind the

basic road pricing model as presented in Figure 1.

Figure 1 Optimal taxation in the presence of social costs from car road travel

with a basic road pricing model

Sources: The presentation of the Figure 1 is a modified version of Walters

(1961) Figure 1.

Within the figurative representation of the basic road pricing model – that can be basically

applied mainly to congestion related externality calculations – (see Figure 1) the following

assumptions are made: each individual makes one trip, as one person per vehicle, along a

single section of road between a common origin and a common destination. The number of

trips, measured as an hourly flow, is plotted on the horizontal axis. The cost per trip defined

as vehicle operating costs plus the opportunity cost of travel time is plotted on the vertical

axis. As the number of trips Q increases, congestion eventually forces drivers to slow

down, increasing the average cost of a trip C(Q) through time delays. The motorist’s

marginal cost coincides with average cost when each motorist accounts for a negligible

fraction of flow and all trips are identical in cost. Hence, the function C(Q) is assumed to

represents both the average cost and the private marginal cost of car (and road) use.

Demand for trips (and trip utility) is described by a conventional downward-sloping inverse

demand curve p(Q). Without road pricing equilibrium arises at the point of intersection G,

where Q

trips are made, each at a cost C

. The equilibrium is inefficient because car users

ignore the fact that they absorb road capacity and therefore leaving less space to the

remaining road users. The excess demand of road capacity imposes time delays and

therefore causes additional time cost to other motorists. Moreover, the private cost curve

C(Q) does not consider additional costs imposed by the individual car user on the society as

a whole, i.e. environmental damage, noise, etc. (see Table 1). The total social cost of Q

trips is TC(Q)=C(Q)*Q, and the marginal social cost of a trip is

MSC(Q)=∂TC(Q)/∂Q=C(Q)+∂C(Q)/∂Q*Q. The social optimum is therefore represented by

the point D of intersection between the MSC(Q) and p(Q). The socially optimal number of

trips Q

is less than Q

. To reach the social optimum road users must bear the cost of C

including the cost they impose on other road users as well as on the general public. This can

be realized by introducing an extra charge for road use at the level of

=MSC(Q

)−C(Q

)=∂C(Q

)/∂Q

, where τ

equals the marginal external congestion

of a car journey and can be extended to include other social costs generated from car use.

is also known in the literature as the “Pigouvian tax” or the “Pigouvian toll” and its

formula is intuitive since it equals the marginal delay imposed by a driver on each other

driver ∂C(Q

)/∂Q

multiplied by the number of other motorists Q

. The social welfare gain

from imposing the additional road use charge is measured by the increase in social surplus

covered by the area DFG. It results from the reduction in total costs minus the reduction in

total benefits due to the decrease in traffic.

The argumentation behind the model described in Figure 1 might be convincing,

nevertheless the model implies a number of rather unrealistic assumptions and difficulties

when it comes to the real world implementation of its formalized mechanism.

Some of the often criticized assumptions introduced into first best solution models for the

sake of its simplification are:

Ubiquity of user charges,

A single road connecting one origin to one destination (no junctions),

One user per vehicle,

Travelers are assumed to have perfect information (no uncertainties),

Vehicles contribute equally to congestion,

Identical individuals except for their reservation price to make a trip,

Traffic flow, speed and density are uniform along the road, and are independent of time,

Congestion is the only market failure; i.e. there are no other transport externalities or

distortions elsewhere in the economy and they would have to be additionally

incorporated into the model framework, and

There are no shocks due to accidents, bad weather, special events, etc.

Rest of the economy operates under first best conditions

Moreover, not only the simplifying assumptions have been subject to critical discussion

among scholars in the field of economics as well as engineering, the modeling framework

and the theoretical idea behind road pricing has been bearing an important number of

substantial concerns. Firstly, imposing road pricing increases drivers’ private costs since it

follows the objective of road capacity management in terms of car travel reduction. The

road pricing revenue accrues to the pricing operator, which is usually assumed to be a

government agency. Hence, after paying the additional road use charge motorists end up

worse off in the first place. The Q

users who continue to travel per unit of time suffer a

cost increase of C

-C

, and the Q

-Q

motorists who give up road use suffer a loss of

surplus that ranges from zero for the marginal user at Q

in the pre-policy-intervention

situation to C

-C

for the new marginal user at Q

. This argumentation is limited to the

road and car user society, if the analysis focuses on congestion externalities. However, the

argumentation can be extended to the society as a whole, including its non-motorised

members through the consideration of social marginal costs of externalities.

Undoubtedly, these losses give grounds for opposition to road charging. The welfare

decline can be compensated only if the government uses at least part of the revenue to

invest into road capacity expansion, to improve alternative transportation means, to cut

other user charges, or to introduce rebates to drivers in some lump-sum manner. In his

discussion of external economies Pigou (1920) implicitly promoted the use of appropriation

suggesting that the revenue from the social cost internalization measure should be “devoted

exclusively to the execution of new and specific road improvements”, to the effect that “in

the main, the motorist does not pay for the damage he does to the ordinary roads, but

obtains in return for his payment an additional service useful to him rather than to the

general public” (Webb and Webb, The King's Highway, p. 250 in Pigou 1920 p. 193). This

argumentation refers to some extent to the idea that the internalisation of external effects

from road use is indirectly linked through road damage to the aspect of infrastructure

network financing.

Secondly, the mere collection of road charging entails infrastructure, operating and

administration costs, including the inconvenience for motorists. Is the demand for car travel

rather price inelastic – as it is generally the case in the short run – the charging revenue

(covered under ADEB in Figure 1) is substantial compared to the welfare gain DFG. But it

should be kept in mind that the net social benefit from road pricing can easily turn negative

as soon as its proportional collection cost per unit of collected revenue becomes high.

Hence, Pigou’s idea of government intervention through internalisation of externalities was

not fully uncontroversial on the basis of regulatory policy principles. According to Knight

(1924) and against the Marshallian tradition of external economies, government

intervention and public transport management were less efficient than private transport

management and were therefore considered superfluous. Almost half of a century later

Coase (1960) brought into the debate the aspect of transaction costs, being in general higher

for multilateral transactions when private parties are involved than for government

intervention. In this case public regulations may be more efficient.

To overcome some of the simplifications and shortcomings of the theoretical approach as

presented in Figure 1, especially with regard to congestion charge modelling, additional

specifications were included in the time-independent model to make it time-dependent.

This is in line with the fact that travel demand is not constant over time. The additions refer

to how travel demand depends on time and how traffic flows change over time and space.

The aspect of special travel demand distribution and road capacity use results from the fact

that road infrastructure supply can be considered as flexible (and extendible) in the long

run, in the short run or at a given point in time it has a limited capacity. For the theoretical

consideration of excess road use demand and congestion (externalities), the objective will

be to model the welfare optimal level of congestion rather than to model congestion-free

social optimum. Nevertheless, first best approaches to road pricing calculations remain

subject to strong criticism. Significant barriers regarding the implementation or the

calculation of the mathematical solution for the different first best optimization approaches

are the unrealistic assumptions underlying the models as well as the estimation of the C(Q)

and p(Q) curves.

With the introduction of second best approaches to marginal social costs calculations at

least the dilemma with some of the difficult to meet assumptions underlying the first best

models have been tackled. According to the economic principles underlying first best road

pricing solutions tolls are calculated equivalent to the external costs generated by each

motorist or traveller. The economically efficient first best solution is assumed to encompass

the optimal road use at its maximum efficiency. Nevertheless, first best pricing has been

criticized as of largely restricted practical relevance turning interest to second best pricing

as being closer to the practical reality. In second best approaches to road pricing

calculations, one or more of the restrictions or model assumptions required for the first best

solution as enumerated above are modified or removed. Beckmann et al. (1956) argue for

example that it is straightforward to compute first best tolls on a road network assuming the

ubiquity of user charges and perfect information, but if tolls are restricted to only some

links of the network or are held constant over time the second best approach becomes more

The consideration of individual time values plays a particular role when road charging is applied given the

objective of a congestion management mechanism. With regard to equity concerns it has been a general

argument whether congestion charges work in a regressive manner, i.e. in favour of high income groups who

are assumed to have higher values of time and therefore their profit from congestion reduction exceeds the

burden of paying the congestion charge (Richardson, 1974; Evans, 1992; Arnott et al., 1994; Small, 1983;

Transek, 2002).

suited for solving the problem. Some examples of the second best based pricing solutions

are cordon pricing around city centres instead of overall network charging, the

implementation of stepwise pricing instead of smoothly time varying pricing schemes, or

road pricing according to fixed instead of varying daily traffic conditions. Thus, “in reality,

the optimal prices are non-unique and deviate from marginal external costs, since they can

be practically imposed only on a set of links, and/or include the effect of several other

restrictions and market distortions, hence, yielding ‘second-best’ pricing settings” (Tsekeris

and Voss, 2009).

In other words, the approximation of the theoretical model with its optimum calculations to

the real life situation requires the relaxation of some of the Paretian optimum conditions.

This is where the optimisation approach is being transferred from a first best to a second

best solution. As it is known the attainment of a Paretian optimum requires the

simultaneous fulfilment of all the specified optimum conditions. The general theorem of the

second best indicates that if one of the Paretian optimum conditions cannot be met due to

some constraints a second best optimum solution is achieved only by giving up all other

optimum conditions. Hence, the failure of fulfilment of one of the optimum conditions

makes the attainment of the remaining conditions obsolete, even if they are still attainable.

The resulting second best optimum situation is therefore by definition attained subject to a

constraint that hinders the realisation of a (first best) Pareto optimum situation (Lipsey and

Lancaster, 1956).

Nevertheless, to solve mathematically for the optimum at point D and for the optimal

pricing level at τ

under first best or second best optimality conditions remains rather

complex. Even the rules necessary for second best optimum attainment are quite

challenging since they have to take into account an array of indirect effects. The

computation of the optimum requires the estimation of the (individual) demand and cost

functions, including elasticity parameters. Even with the current advances in computation

techniques and data collection methods this remains a demanding undertaking, regarding

the requirements and theoretical assumptions inherent in marginal social cost and efficient

price calculations (Lindsey, 2003; Lindsey and Verhoef, 2001; Nash, 2001; Rouwendal and

Verhoef, 2006; Rothengatter, 2003).

They involve detailed information on the composition

and the sum of the external costs caused by transport, including e.g. heterogeneous trip-

time preferences and time values of individual road users, i.e. anonymous vs. non-

anonymous, or type-specific pricing. Especially the estimation of overall congestion costs

and the money value of lost travel time based on opportunity costs calculations face major

theoretical and measurement obstacles (Button, 1993b).

Rothengatter (2003) points to additional restrictions implied by the simple textbook

approach that throw a critical light on the implementation of the marginal social cost

pricing approach:

Complex measurement,

Equity is ignored,

Dynamic effects including investment decisions and technology choices are not taken

into account,

Financing issues are ignored,

Institutional issues (public economics) are ignored,

Price distortions elsewhere in the economy are ignored.

2.1.3 Alternative approaches to marginal social cost pricing

Another critical aspect of (social) marginal cost pricing taking into account some of

Rothengatter’s (2003) critical notes is the distinction implied between short run vs. long run

marginal cost. In line with the short run marginal cost pricing approach no additional

infrastructure provision is taken into account rendering infrastructure capital costs

irrelevant. The long run marginal cost pricing approach assumes that infrastructure

experiences an optimal expansion as response to additional traffic, reducing in turn the

additional negative effects from traffic congestion externalities. Hence, if capacity is

optimally adjusted up to the point where the additional capital costs of expanding capacity

For extended literature review on road pricing theory also see Langmyhr (1995 and 1997), Morrison (1986),

and Small (1992).

equal the reduction of the otherwise generated marginal costs from excess travel demand

the short run marginal cost and long run marginal cost equalise. The main problem with

equivalence assumption between the short run and long run marginal (social) cost pricing is

the fact that in most cases road capacity cannot be easily and flexibly, optimally adjusted to

demand.

As suggested by the neoclassical economic theory marginal social cost pricing, defined as

the sum of the costs imposed by an additional user on the network, on other users, and on

the society as the whole, is the welfare maximising, first best pricing scheme for charging

transport infrastructure use. It is also general consensus that short run marginal cost is the

proper pricing approach. Nevertheless, the considerations of long run marginal cost pricing

methods may be appropriate when decisions about the provision of additional road capacity

need to be accounted for within the optimal price calculation. This is for example the case

when self- financing constraints of the network or an inter-modal transportation system are

applied to the calculation of the marginal social cost optimum. The optimal road charge

will then tend to exceed the intercept between the social marginal cost and the travel

demand curve and be located above it, where the average cost curve and the car use

demand curve cross each other.

However, on the theoretical basis and under certain assumptions Mohring and Harwitz

(1962) accomplished to bring together the short run social marginal cost pricing approach

and the postulate of infrastructure capacity coverage. The resulting cost-recovery theorem

implies that the revenues from short run social marginal cost pricing suffice to pay for

optimal capacity if capacity is perfectly divisible and supplied at constant marginal cost,

and user costs are homogeneous of degree zero in usage and capacity. Therefore, the cost-

recovery theorem seems to overcome the conflict between short run social marginal cost

pricing and the average or the full cost pricing approach based on revenue coverage

criterions. Part of the research that followed the idea of road pricing serving as a financing

source to pay for (road) infrastructure provision explored the robustness of the cost-

recovery theorem to relaxation of assumptions (Newbery, 1988 and 1989; Small and

Winston 1988; Small, Winston and Evans, 1989). Again, already Dupuit (1962 [1849])

referred to road charging as to an instrument for covering long-run costs of road

construction and maintenance in the sense of a “funding toll”, rather than to manage road

use in terms of a “decongestion toll”.

Especially growing political interest in and willingness to actually implement road pricing

schemes in practice helped to navigate the theoretical debate about second best marginal

social cost based price calculations for road use towards alternative pricing theories, where

the focus also lies on balanced budgets within as well as between the transportation sectors,

similar to the idea of the cost recovery argument. To mention just two alternative pricing

rules, average cost pricing and Ramsey (1928) pricing are both considered as deviations

from marginal social cost pricing. According to the average cost pricing rule prices are

equal to the sum of financial costs of the mode in consideration divided by its total volume.

The structure of resource costs, i.e., fixed vs. variable, sunk or not, etc., and the type of the

transported medium, i.e., goods vs. passengers, are not specifically taken into consideration

within the calculation approach. The main objective of average cost pricing is cost

recovery. Many forms of average cost pricing exist since the numerator and the

denominator are to some extent arbitrary. Therefore, several volume indicators can be used

for the cost calculation, e.g., trips or vehicle kilometres for passengers and ton kilometres or

vehicle kilometres for freight. Furthermore, differences within accounting rules used within

the average cost calculation are not always uniform, e.g., in terms of depreciation rules, etc.

and can therefore result in different total cost concepts (Jha, 1998).

The discussion about the implementation of short run marginal social cost pricing to induce

“efficient” usage of roads vs. average cost pricing to finance them goes back to the research

era of Dupuit and Pigou. The difference between the two approaches lies especially in the

requirements related to their calculation, where marginal social cost pricing requires the

variation of charges with respect to space, time and vehicle, or user characteristics calling

for finer pricing instruments than average cost pricing.

The rule applied for the calculation of Ramsey social cost prices is based on the idea that

prices are set as optimal deviations from marginal social costs. The deviations are required

to meet cost recovery targets for the transport sector as a whole. In the case when the

revenues collected based on the marginal social cost pricing do not cover the financial cost

of infrastructure (use), Ramsey pricing requires that the margins (price-marginal social

cost) are increased in a way that is inversely proportional to the elasticity of demand in the

relevant market (Proost and Van Dender, 2003).

In economic terms, the basic idea behind the emerging alternative cost calculation rules is

to calculate the optimal (most efficient) deviation from the first best to the second best

pricing solution. This may also include the calculation of “third best” pricing, i.e., setting

“quasi” first best tolls as if second best distortions do not exist or the application of other

“rules of the thumb” to make the pricing calculation more appropriate for implementation

in practice. Variable pricing schemes can still be recognised as generalizations of the

marginal social cost pricing approach. The advantage of variable pricing rules is that they

bring about efficient behaviour in the Pareto equilibrium without knowing initially what the

state of efficient behaviour is supposed to be.

2.1.4 Policy frameworks and road pricing calculations

The implementation of road pricing policies and the application of transparent and

politically approved calculation approaches can be facilitated by the introduction of

regulative policy frameworks.

Therefore, the implementation of market-based instruments for internalisation of external

cost has been validate in EU Directives, particularly related to infrastructure cost pricing.

Hence, according to the amendment of the Directive 1999/62/EC adopted on 27 March

2006 – better known as the Eurovignette Directive on road charges – the European Union

allows member states to levy tolls on all roads. Regardless its constraints, the Eurovignette

Directive is a substantial leap forward towards the implementation of a European road

charging policy (EC 1999, 2008, 2006, and 2009).

An important constraints of the Eurovignette Directive is the requirement that revenues may not exceed

related infrastructure costs. Furthermore, the Directive limits the differentiation of charges according to

capacity or environmental criteria allowing a mark up of maximum 25 % only for mountainous areas to

reflect the higher infrastructure costs.

Concerning the question about the methodological approach to calculate or estimate cost

components or price elasticities as the relevant inputs for the application of the economic

theoretical concept of marginal social cost pricing, the European Commission has raised the

issue of internalisation in several strategy papers, such as the Green Book or the Green

Paper ‘Towards Fair and Efficient Pricing in Transport’ (CEC, 1995), the White Paper on

efficient use of Infrastructure, the European Transport Policy 2010 (CEC, 2001) and it’s

midterm review of 2006 (EC, 2006), including a number of research projects. The EC

White paper of the overall transport strategy and the midterm review emphasize the need of

fair and efficient pricing considering external costs. As part of the 4th Framework

Programme and the following

, the Commission sponsored a large amount of research on

how to implement pricing policies, in terms of feasibility and acceptability problems and on

resulting implications from their implementation (Nash, 2001).

A great number of the studies carried out on the behalf of the EU are concerned with the

evaluation and the quantification of external costs:

EU-Research projects of several framework programmes to estimate external costs

(such as UNITE, ExternE, GRACE, etc.),

Other EU projects on external and Infrastructure costs, particularly marginal costs of

Infrastructure use – towards a simplified approach (CE Delft, 2004),

National research projects and studies on external costs (particularly for the UK, the

Netherlands, Switzerland, Austria, Germany),

International estimates of external costs,

EU-proposals to standardize marginal cost estimation (High level group approaches),

EU-Networking projects to discuss pricing instruments (CAPRI, IMPRINT, MC-

ICAM).

A number of EU research projects has been undertaken to evaluate and draw conclusions

on potentials of efficient transport pricing. As summarized by Sikow-Magny (2003) the EU

projects TRENEN-II-STRAN, PATS and AFFORD investigated possible

Detailed description of funded EU Framework Programmes can be found under “Find Funding”, under

http://ec.europa.eu/research/index.cfm (25.12.2009).

operationalisation options of the marginal social cost pricing principle. The projects

PROGRESS, CUPID, DESIRE, and MC-ICAM dealt with implementation possibilities of

marginal cost pricing.

Also the implementation and evaluation of existing (road use) pricing policies on the

national level are of great importance for the extension and a better understanding of the

pricing mechanisms. In Europe, most often national pricing strategies were introduced in

the sector of heavy goods vehicle (HGV) charging including several countries, such as an

HGV-fee mostly covering the intercity national road network (Autobahn) in Switzerland

since January 2001, Austria since January 2004, or Germany since January 2005. Urban

road pricing schemes, mostly aiming at the reduction or a better management of inner-city

car traffic in terms of a congestion charge, were introduced for example in Bergen

(Norway), London (United Kingdom), and Stockholm (Sweden), as well as in Valetta

(Malta) and Milan (Italy).

The advantages of road pricing are the flexibility of implementation, the direct reference to

car use instead of to its inputs, and the ability to differentiate it between varying user

categories. The implementation of marginal social road pricing in practice is likewise

concerned with objectives of welfare distribution and social equity, especially when is

discussed in practice focusing on public acceptance concerns.

Historically, the main fiscal instrument in practice affecting motor vehicle use has been the

fuel tax. Its main purpose has been to raise revenue rather than to finance infrastructure

investment. The mechanism behind road infrastructure finance in Germany is not based on

a clear-cut (tax) revenue inflow and finance outflow account, i.e., not all of the finance

outflows are covered from inflows collected for the exclusive purpose of road infrastructure

investment but from other state funds. Due to the growing gap between rising motorised

travel demand and limited budgetary sources provision of road capacity has become a

major challenge. One reason for the expected melting away of traditional, primarily tax

For an overview of EU research projects on potentials of efficient transport pricing see

http://www.transport-pricing.net/ (01.11.2009).

An overview of implemented road pricing schemes can be found under

(http://portal.wko.at/wk/format_detail.wk?angid=1&stid=240298&dstid=7164&opennavid=31614

(01.11.2009).

based road infrastructure financing sources in the future is the likely revenue shortfall from

fuel taxes entailed be fuel consumption efficiency gains. While technological

improvements of fuel use efficiency lead to decreasing fuel consumption, car use and

therefore infrastructure use stagnate or even continue to grow. Within the current EU

context road charging has been also regarded as a more and more important revenue source

for the construction of the trans-European transport networks and 30 additional

transnational projects declared of particular interest to the European Union.

Therefore one

objective mentioned in the White Paper of 2001 (CEC, 2001) states “The thrust of

Community action should be to replace gradually existing transport system taxes with more

effective instruments for integrating infrastructure costs and external costs (in the price of

transport). These instruments are, firstly, charging for infrastructure use, which is a

particularly effective mean of managing congestion and reducing other environmental

impacts, and, secondly, fuel tax, which lends itself well to controlling carbon dioxide

emissions”.

2.1.5 Introduction of road charging in Germany

In Germany freight or heavy goods vehicles (HGV) have been charged a distance

dependent toll for the use of the intercity national road network (Autobahn) since January

2005. So far charging of cars remained on the level of a political debate and feasibility

studies. It is probably not unrealistic to assume that the existing HGV tolling technology

could be extended to the sector of passenger cars as well as urban roads after advanced

technological adjustments have been undertaken.

In general, provision of public roads together with government intervention through the

introduction of road charging can be considered also in Germany an efficient alternative to

revenue seeking for infrastructure investment and traffic management. Often, practical

Decision 884/2004/EC.

According to the RECORDIT program the difference between external costs in interurban and urban road

use varies depending on vehicle type from 5 to 60 Euro-Cents per km.

implementations of road pricing function rather as revenue generating mechanisms, than as

road use demand management or restrain tools (Paulley, 2002).

The implementation of road charging policies requires the specification of the level and the

type of charging. In this work it is assumed that imposing a road charge will align the

private cost of a car travel closer with the marginal social cost of car use. For the model

implementation described in Chapter 5 the level of road charging is set based on social

marginal cost pricing calculations for Germany and with regard to the results obtained for

other EU countries. The scenario analysis is carried out based on 0.05 Euro per km,

distance dependent road charge imposed as a mark-up on the variable car travel costs on

private car drivers. In line with the economic theory the charging measure applies to travel

along a link, covering the entire road network. Hence, there is no incentive for road users to

divert their traffic to uncharged roads. The 0.05 Euro per km road charge rate is a lower

bound, averaged estimate drawn from a survey of German as well as European studies on

road infrastructure cost assessment as well as external average and marginal social cost

calculations. However, methods used for (full) road transport cost assessment vary and

resulting numerical estimates differ since they are often based on different assumptions as

to the kind of costs included into the calculation. The rate of 0.05 Euro is set as one half of

the reference value from average external cost calculations for cars and is assumed to

correspond to the lower bound from marginal external cost calculations taking into account

the general external costs categories (see Table 1) (Herry and Sedlacek, 2003; IMPACT,

2008; IVT, 2004; Infras/ IWW, 2000 and 2004; RECORDIT; UNITE).

Albeit the historical tradition and established and uncontroversial theoretical reasoning –

especially among economists – in favour of road charging, the pricing instrument has been

Even though road charging schemes (when based upon Pigouvian taxation principles) are often related to

the principles of internalization of external costs from motorized road use, they not always meet this maxim.

Verhoef (1995) suggests therefore, the need to assess externalities occurring after implementation of

“optimal” road charging policies. Despite the difficulty to determine prices for positive or negative

externalities of a different spatial scope (Owens, 1995), this aspect is somehow relevant for the equity and

distributive impact evaluation in the context of road use charging and revenue redistribution.

In general, the implementation of a road charge is expected to bring about a more efficient utilization of the

entire transport system and thus create a social surplus. However, the policy definition implied does not fully

correspond to the marginal social cost pricing principles. To some extent, the implementation of the road

charging policy follows the restrictions implied by the modeling framework applied in this work. A more

detailed description of the policy scenario definition for the model implementation is included in Chapter 5.1.

ever since struggling for public acceptance and a broader dissemination in practice (see also

Chapter 1.1.2) (Gaunt et al., 2007). Transport economists often endorse the positive

economic effects from road charging. The research literature often cites the Smeed Report

(U.K. Ministry of Transport, 1964) claiming road charging being irrefutable. Nevertheless,

reluctance towards road pricing measures among politicians and the public is the common

case. Only recently, the implementation of road charging has been gaining interest partially

due to technological innovations and new electronic solutions, which could remedy some of

the existing shortcomings. More sophisticated road charging adjusted to specific time

frames, road categories, or even user groups could be implemented efficiently ensuring

convenience and lower operating costs, also taking into account the controversial issue of

data confidentially.

This implies the importance of investigating the interrelation between acceptance of road

charging measures, its welfare and distributive effect, and the underlying sociodemographic

and economic characteristics, land use structure, territorial population distribution within an

overall macroeconomic picture. Oberholzer-Gee and Weck-Hannemann (2002) suggest the

incorporation of environmental quality within the design of road charging instruments to

ensure the fairness of the system and in this way to win public acceptance. Viegas (2001)

discuss the influence of equity on political and public approval of road charging

mechanisms. Hence, even though the overall economic net welfare effect from road pricing

will most likely turn out positive, it may not enhance the public acceptance as long as it

neglects distributional and equity effects concerning the burden and profit sharing of

different socioeconomic groups. However, the expected individual welfare enhancing effect

from the pricing measure can be generally brought about through the road charging revenue

use. The design of the revenue redistribution scheme according to established

technological, institutional and acceptability related barriers could be implemented to

counteract welfare equity concerns. Based on results from standard neoclassical methods of

welfare analysis and from equity assessment (Atkinson, 1970) simulation analyses of net

benefit and equity implications can be carried out introducing different road charging

revenue redistribution schemes between the general consumer sectors included in the

model. Hence, the policy simulation studies varying in revenue redistribution are described

in detailed in Chapters 5.1 and 5.2. Chapter 5.3 presents the results from the policy reform

implementation within the modelling framework applied in this work. The results provide

valuable implications for potential acceptability of such policy reforms. The modelling

results for Germany lead to the conclusion that differences in distributional effects with

respect to equity impacts are strongly related to the revenue recycling patterns as well as to

household specific travel demand profiles that are depending on income and residential

location.

2.2 Discussion: acceptability of road charging

Despite the (overall social) welfare improving function of road charging, hardly anybody

imposed to such a measure values it as putting the society better off. This negative

individual perception is one reason for the general lack of public acceptance of road

charging. Also the aspect of public revenue generation is generally ignored. Another reason

inhibiting the introduction of road charging comes from uncertainty about the negative

equity impacts resulting from the measure. In many cases road charging is suggested to

have a regressive impact (with regard to income), which means it is more disadvantageous

for the poorer drivers. Therefore, even though in many countries the existence of

institutional barriers is blamed for impeding the realisation of road charging schemes

(Glazer et al., 2001; Schade and Schlag, 2003), the absence of public and of political

acceptability remains the most difficult barrier to eliminate on the path to the introduction

of road charging measures (Bartley, 1995; Luk and Chung, 1997; Jones, 1998 and 2003;

Schade and Schlag, 2000a, 2000b, and 2003; Link and Polak, 2001; Jaensirisak et al.,

2005). The examination of the acceptability of urban road pricing conducted in eight

European cities within the European project PRIMA (2000) came to the conclusion that on

average less than 30 % of the population support such measures. Consequently, only few

implementations of road charging exist this far, of which some are suboptimal solutions of

indirect charging and do not take into account the distance driven, hence the individual road

use (e.g. cordon pricing, city toll, etc.). Except for urban road charging schemes established

in Singapore, Oslo, Bergen, Trondheim, Stavanger, or London (often referred to as

congestion charging), sophisticated systems of this kind have never reached beyond the

planning stage. Gaunt et al. (2007) lists numerous cases, where intentions of implementing

road charging never materialized. Yet, the set-up of road charging schemes has not only to

satisfy public acceptance criteria, but it has to be effective in meeting its primary objectives

of road use management, or revenue generation (Jones, 1998). The significance of public

acceptance for an unobstructed functioning of road use charging is equally valued as and

also dependent upon the technical success of the scheme. Also important is the integrated

transport policy accompanying the pricing reform, e.g. in terms of the level of car

ownership and use related taxes and levies other than road charging (Gray and Begg, 2001).

Obviously, public acceptance of road charging depends on the amount charged, where

lower charges are more likely to be accepted than higher ones (Safirova et al., 2002 and

2003; Jaensirisak et al., 2005).

However, public resistance to road use charging is neither a static nor an irreversible

phenomenon. It is linked to numerous factors, such as sociodemographics, economics,

residential geography, individual attitudes, political views, or the design together with the

implementation characteristics of the charging scheme itself. Car ownership or car

availability and perception of benefits to oneself and society have also a significant

influence on the success of the implementation of road charging policies (Jaensirisak et al.,

2003). Studies evaluating efficient pricing policies suggest, that pricing policy design and

implementation need to consider in the first place resistance from car users. For a road

charging policy to be publicly approved it has to be perceived as beneficial to each

individual who is subject to the measure, to the overall public, or to both. Is the measure

putting the society and its members better off, increasing their welfare, or environmental

quality, it is more likely to be perceived as worth being paid for and as a result gains

acceptance (Giuliano, 1994; Goodwin, 1997). According to Schlag and Schade (2000a) one

of the most decisive factors in favour or against road charging reforms is the way in which

the revenues from the measure are reallocated. Therefore, hypothecating revenues towards

specific targets, such as e.g. public transport, increases significantly public support for road

user charging reforms.

Another crucial factor related to the acceptance of road charging policies are equity

concerns linked to distributional effects (Gaunt et al., 2007). Although it is known that road

pricing will normally generate a net welfare surplus the argument alone may not be

particularly convincing in a real-world situation. The reason is that the argument neglects

equity effects in terms of the distribution of costs and benefits of the measure across

socioeconomic groups. Equity and distributional impacts can embrace the social or

geographic sphere, and have an adverse effect on population with lower income or

disadvantaged social status or the economic prosperity of certain spatial zones. Putting the

needy groups relatively worse off than others or than was their initial state before road

charging, increases the chance for the measure to be rejected (Jones, 1998 and 1991; PATS

Consortium, 2001; EURoPrice, 2002; Bell et al., 2004). A transparent and comprehensible

design of the policy measure and the revenue use are likely to evoke public acceptance. The

design of the revenue reallocation scheme can help to communicate the overall benefits

from the measure making them more visible.

Putting the two aspects together, adequate compensation of disproportionately “penalized”

entities by the road charge through the design of the revenue reallocation scheme affects the

resulting welfare distribution and finally the acceptance and feasibility of the policy

measure. As a number of studies concerned with acceptance of and equity effects from road

charging reforms acknowledge, that revenue reallocation schemes designed to cut other

taxes or improve the public transportation system contribute positively to the endorsement

of road user charging (Jones 1991; Ison 2004; Harrington et al. 2001). The design of

revenue redistribution schemes that will have a positive welfare outcome and lead to public

support of road charging policies requires information on potential impacts from road

charging on different population groups in the first place. This overlaps with the objective

of this thesis work. There exist a number of theoretical studies concerning equity effects

(Arnott et al., 1994; Glazer and Niskanen, 2000; Richardson, 1974; Richardson and Chang-

Hee, 1998; Small, 1983). They all largely conclude that equity effects will in general

depend on the design of the road charging policy including revenue recycling and on

socioeconomic differences in travel patterns, i.e. how mode and destination choices differ

across income or other population groups. However, there are significantly fewer studies on

detailed, quantitative assessments on equity impacts. The objective of this dissertation work

is to contribute in closing this knowledge gap using an adequate methodological approach

and an extensive data basis for Germany. Chapter 2.3 sheds more light on the theoretical

discussion about equity, welfare and distributive issues that arise in the context of road

charging.

2.3 Equity concepts and criterions

The design of the charging policy as to the level of charges imposed on each road user and

the revenue redistribution scheme are two important aspects of equity and redistribution

considerations in the context of road charging policies.

In general, not all impacts induced by imposition of charges on road use can be perfectly

assessed. Ex-ante assumptions for supply adjustment to changes in demand are not

straightforward to make. Potential elasticity parameters gathered from different empirical

studies and literature sources are afflicted with an array of uncertainties. Implications for

road charging scheme designs, as to who is charged how much, where and at what time, are

derived from for the most part simplistic, model based studies. Hence, it needs to be kept in

mind when analysing impacts and discussing scheme designs of road charging and revenue

recycling that the theory behind it is founded upon reduced assumptions compared to the

reality (Pigou, 1920). However, when it comes to welfare impacts, high-income categories

are most often recognised as ending up better off after road charging implementation than

low-income groups (Else, 1986; Giuliano, 1992; Hau, 1992). With everyone paying the

same amount of road use charge for a distance unit travelled (km), disproportional burden

is likely to be levied on low-income car users. This aspect may be seen as controversial,

nevertheless it lays down the link between road charging and economic concepts of welfare

distribution, equity and acceptance.

Realised mobility patterns and travel expenditure profiles of individuals or private

households reflect their choices made based on subjectively valued choice alternatives,

taking into consideration the utility and the cost of the choice alternative. Travel cost

increase through the imposition of road use charges or augmented fuel taxes introduced for

example as measures to internalise the social marginal cost of negative externalities from

motorised vehicle travel. They will generate (fiscal) revenue, but at the same time alter

travel choice behaviour.

Social marginal cost based road charging generates a social welfare gain, which can be

redistributed to the different economic sectors or groups within the society. Distributional

and equity concerns entail the question to which aggregates and in what division the

revenue should be reallocated. Aggregates such as car users or people belonging to the

same income category still constitute a heterogeneous group of individuals. Transferring

revenues from road charging according to households’ income level might still miss the

equity criterion in terms of welfare distribution (Morrison, 1986, p.93).

The concept of distributive equity is difficult to grasp from the economic theory

perspective. Distributive effects, equity effects, equity distribution, welfare effects, welfare

economics are some of the terms related to the effort to assess the “fairness” or “justice” of

a policy intervention in the functioning of markets or the society.

In economics the concept of equity or the idea of fairness originates from taxation or

welfare economics, where it is distinguished from economic efficiency in overall evaluation

of social welfare. Describing welfare distribution the equity concept is used to describe a

fair and socially acceptable allocation in contrast to economic inequality. However, equity

is used in a broader sense. In public finance horizontal equity describes the idea that

individual ability to pay taxes should determine the tax amount paid and this applies to all

individuals. Hence, same individual affordability means same tax burden and is in line with

the concept of tax neutrality and against distortion of (economic) behaviour. On the other

side, the concept of vertical equity used in public finance refers to the idea that people with

a greater financial ability should pay higher taxes. In contrast to a tax burden proportional

In established literature on equity distribution the terms equity, fairness, or justice are used as synonyms

(Blanchard, 1986).

to income stands the concept of progressive taxation, where with increasing income an

increasing proportion of it has to be allocated as tax contribution to the state budget.

Progressive tax policy is further associated with distributive justice and redistribution.

Hence, the total consumer welfare is not the only aspect to be examined. The way

consumer welfare is distributed among society’s members and how equitable that

distribution is are at least as relevant. The general consideration of equity with respect to

horizontal vs. vertical equity can be linked to Rawlsian principles of justice (Rawls, 1999).

Hence, horizontal equity refers to the distribution of welfare among individuals who are

otherwise identical, drawing from Rawls’s “principle of equal opportunity”. Vertical equity

refers to the distribution of welfare among individuals who are unequal in other respects,

drawing from Rawls’s “principle of difference”. Additionally, a third form of spatial equity

has been specified by Raux and Souche (2004) implementing Rawls’s “principle of liberty”

as the right of access to any location in space.

Imposition of road charges on road users can be understood in a broader sense as a form of

policy intervention. The difficulty with the assessment of fairness and equity is that it is in

principle a normative criterion – the question is, can a universally accepted definition of

fairness exist and be operationalised in terms of measurement. Distributive effects in terms

of welfare reallocation between different population aggregates after policy reform

implementation are possible to calculate. It is not an exception that policy evaluation

integrates distributional considerations. To conclude, whether the observed shifts of

burdens and benefits are consistent with equity conditions is much more ambiguous.

Firstly, equity does not have to mean that everyone ought to bear equal gains or losses.

Secondly, accounting for equity effects does not automatically mean, that equity concerns

on all sides are equally taken into account. Defining and ensuring fairness is all too often an

impasse rather than a straightforward decision.

There exist a number of partially even inconsistent concepts for equity (Blanchard, 1986).

Langmyr (1997) takes on from the literature “thirteen distributive principles or criteria”

pointing to the relevance of the specific environment to which they are applied. Examples

for such analytical dichotomies proposed to assess equity distributive concerns are, e.g.

horizontal versus vertical equity, as being probably the most common approach, telelogical

versus deontological ethics, or equality of chances versus equality of outcomes. Blanchard

(1986) emphasizes seven rules to measure equity or fairness of policies, after exhaustively

illustrating recent attempts in literature to prove different conceptions of justice in

theoretical terms. He refers to such fundamental works as “A Theory of Justice” by Rawls

(1971), equity norms as defined by Hochschild (1981), Levy et al. (1974), Ryan (1981),

Miller (1976), or Lucy (1981).

Possible criterions for classifying or evaluating equity definitions are: explicitness, scope,

or to what systematic extent they interrelate. Hochschild (1981) systematised the following

equity norms: strict equality, need, effort made, money invested, results, ascription, and

procedure. Overall equal shares are recovered by “strict equality”. “Need” and “effort”

imply that redistribution takes place relative to given needs and expended efforts,

respectively. An analogous implication applies to the equity norms “invested money” and

“product”. If benefit or burden distribution is based on genetic or socially defined

criterions, e.g. age, sex, or income, Hochschild (1981) talks about “ascription” being the

underlying equity norm. Finally “procedure” links the distribution scheme to some specific

procedural approach such as a certain kind of ordering or random lottery (Hochschild,

1981). Some of the equity concepts defined by Hochschild (1981) can be possibly applied

for the impact assessment from car road charging; they can be used as guidelines for

designing and evaluating revenue collection and redistribution schemes under equity

considerations. However, the application of these norms depends on the formulation and

the scope of the research question and is likely to be limited by the availability of data or

information required for the assessment of these equity norms.

Other equity concepts found in the literature are “market equity”, “equal opportunity”, and

“equal results” formulated by Levy et al. (1974). The concepts constitute a continuum with

respect to the redistribution degree going from no redistribution when “market equity” is

assumed to significant redistribution when the equity objective are “equal results”.

Regarding the definitions underlying each of the three fairness concepts, differences as to

the nature of the distributed or redistributed benefit need to be taken into account.

Distinction is therefore made between the quality and the monetary value of a service (-

unit) that is being reallocated and the quality or scope to which the service has met the

underlying objective. Applying this delimitation to the case of road charging revenue

collection and redistribution, this would imply the distinction between provision of (road)

infrastructure vs. assuring accessibility or meeting mobility demand financed by revenue

from introduction of the policy measure. The latter concept corresponds better to the equity

norm of “equal results”. Hence, “market equity” is achieved when the benefit redistributed

to the individuals corresponds to the unique share provided at the cost of the benefit. This

happens disregarding pre-existing inequalities in income or wealth distribution, which in

fact are the prior basis for the distribution of the burdens and then for the allocation of the

benefits. The “equal opportunity” principle implies that benefits are distributed at equal

values irrespective the individual contributions made to the cost of the benefit provision.

This depicts then a case of true redistribution, when the service is financed by a progressive

scheme, meaning those better off contribute higher cost shares to finance it, but everybody

is provided with equal shares of the service. Finally, fairness can be achieved when “equal

results” are obtained. To translate this standard to the case of revenue redistribution from

road user charging it could mean that after augmenting the price of private car use, revenue

from the price increase is used to provide each individual with better or at least equal

mobility opportunities with regard to the initial accessibility or the availability of “travel

tools”.

Although existent definitions of equity norms originating from different “schools” are in

some way comparable, their systematisation with regard to underlying implications can

make a considerable distinction. The critical point is, how far does the equity norm refer to

the allocation of resources vs. the allocation of outcomes. Is the road charging revenue

distributed to equal monetary shares per car user, or is the objective of the road charging

revenue redistribution to assure that each individual can reach his place of work and make

the trip back home bearing an acceptable cost. It seems a reasonable assumption however,

that the allocation of outcomes is determined by an initial distribution of resources. Ryan

(1981) stresses the equity concepts “distribution of resources” (calling it “fair play”) vs.

“distribution of results” (terming it “fair shares”) contrasting them against each other within

an ideological context.

Other examples of equity norms are the definitions formulated by Miller (1976). Hence

distribution can arrive at fairness by “right”, “desert”, and “need”. “Right” implies that

legislation ensures equity by entitling members of the society to receive services, benefits,

and alike, irrespective the considerations. Within the concept of “desert” fairness is reached

when benefits are earned and therefore merited. The fairness standard according to “need”

assumes the existence of some basic “subsistence needs”, which have to be satisfied.

Therefore, the subsistence needs define the share of benefit individuals are entitled to in

order to establish fairness.

Lucy (1981) delineates five equity norms taking into consideration positions of practicing

planers and particularly with regard to distribution of “urban services”: “equality”, “need”,

“willingness to pay”, “demand”, and “preference”. New are the concepts “willingness to

pay”, “demand”, and “preference”. Fairness is achieved when services are allocated only to

those demanding the service and based on their individual “willingness to pay”. Therefore,

only these users pay and their “willingness to pay” is determined by their ability to pay.

This conception can again be critically contrasted against Ryan’s (1981) idea of “fair play”

and “fair shares”. Since one’s willingness to pay reflects one’s ability to pay, the initial

allocation of resources determines to some extent his demand for a specific service. It is

likely to assume that one’s ability to pay does not reflect one’s need for the service, when

referring to the norms “equality of results” or “fair shares”. According to the norms

“demand” and “preferences” equity is satisfied when individuals are endowed with services

or benefits they want. Individuals can communicate in different ways what it is they want,

but they are not obliged to. “Demand” and “preferences” also consider the existence of

individual desires that cannot or simply are not communicated or being asked for. The

notion of what people or the society desires introduced into the summary of equity concepts

is somehow different to what has been discussed so far. They basically address what

individuals want rather than what they are entitled to (Lucy, 1981).

Referring the two types of commonly considered spatial (or horizontal) equity and social

(or vertical) equity impacts, marginal social cost charging might be perceived as fair from

the perspective of charging all road users for the externalities associated with their car

travel. At the same time, marginal social road charging might be considered as “socially”

unfair due to the fact that it is harder to bear for low income road users since it absorbs a

relative greater proportion of their income (socially marginal journeys are not equal to

economically marginal journeys). Distance based charging schemes applied to links and

routes might appear more acceptable from a social equity perspective compared to cordon

based charging. Potential existence of (temporally) uncharged links or routes implies a

trade-off in routing and travel time choice depending on the individual value of time of the

road user. Spatial equity refers to the road users’ residential location and travel destination,

concerning in particular cordon or area charging policies. Cordon or area charging is likely

to be regarded as inequitable if it is designed to apply a different charge to persons residing

just within and just outside the boundary, both travelling the same distance within the

boundary. Therefore a link or distance based system as is more likely to be viewed as

favourable under equity concerns in contrast to an area or cordon based scheme.

In general, it has been argued if and to what extent road charges work in a regressive way,

i.e. putting high income groups better off while penalizing population with lower income.

The effect of road and in particular congestion pricing on (vertical) equity has been

emphasised to work in opposite directions. On one hand high-income earners bear a higher

tax burden by paying elevated income taxes as well as taxes and levies related to car use

and car ownership (i.e. fuel or annual vehicle taxes), driving their cars more intensively and

purchasing bigger and more expensive cars. The set-up of the current road finance policy is

another aspect in this line of thought. Assuming the case when infrastructure provision is

financed from taxes, e.g., income taxes, and the implementation of road charging would

provide an alternative source to infrastructure financing, allowing to lower the initial tax

burden, then through the substitution of tax funded infrastructure investment and tax rate

cuts highly taxed income groups would experience a higher tax relief than low income

earners. On the other hand the same population group is assumed to have higher values of

time and therefore its profit from congestion reduction exceeds disproportionately the

burden of paying the congestion or the general road charge.

However, the assumption that individuals with higher incomes attribute higher values to

their time is not straightforward (Jones, 1992). As illustrated by the following example,

people with a high marginal utility of time and of money (e.g. a single parent worker) may

exhibit the same or similar value of time as people with a low marginal utility of both (e.g.

a wealthy holiday maker) (Langmyhr, 1997).

On the contrary to affluent individuals, population groups with small economic margins are

expected to experience higher welfare losses from road charging. Living in the suburban

rather than the city centre area, low income groups face poorer public transport

opportunities. They are also assumed to be less flexible in their time of activity, especially

time of work choice to possibly avoid peak hour charges (Richardson, 1974; Small, 1983;

Cohen 1987;

Evans, 1992; Arnott et al., 1994; Transek, 2002).

However, reversed arguments exist saying that low income groups are likely to be the

winner from road charging, especially when they live in an environment where the car is

not the principal transport mode. Most European regions and almost all European cities are

more or less car-independent where accessibility by public transport, walking or cycling is

satisfactory and therefore a competitive substitute to car mobility. Assuming that less

affluent population groups are more likely to patronise travel modes other than use the car,

they will be less affected by road charging policies.

As a result, road charging will have a

progressive impact (Glazer and Niskanen, 2000). Since in general road charging revenue is

spend on public transport improvement (Evans, 1992), low income groups being frequent

public transport users (often captives) are more likely to profit from the charging car road

use than persons with higher incomes. These contrasting views indicate the difficulty in

finding a clear-cut conclusion about the distributional and equity impact of road charging.

Despite methodological uncertainties, quantitative studies of road and congestion charging carried out for

cities such as San Francisco (Schiller, 1998), Oslo (Fridstrom, 2000), Gothenburg (Transek, 2002), and

Cambridge, Northampton and Bedford (Santos and Rojey, 2004) indicate that high income population groups

experience a greater negative welfare impact than low income groups due to their high car use intensities and

their preference to reside in (suburban) areas with limited access to public transport. Given an equal

redistribution of the road charging revenue those with low incomes would end up being better off (Arnott et

al., 1994).

Even so, case-adequate or sufficiently progressive revenue recycling schemes could ensure

that all income groups benefit overall (Small 1992), nevertheless each road charging policy

needs its specific evaluation.

As Eliasson and Mattsson (2006) conclude from their studies of theoretical literature on

equity and distributional effects on car users following the introduction of infrastructure

pricing policies, equity outcomes will in general strongly depend on the design of the

policy instrument as well as on varying travel patterns from socioeconomic differences of

population groups.

Evaluation of road charging policies needs to consider distributional

effects before and after the implementation of different revenue redistribution schemes,

comparing net welfare surplus with the total distributional effects. Drawing the conclusion

whether the measurable shifts in welfare distribution caused by implementation of the

policy reform are equitable is far from straightforward. Even though different methods for

quantifying equity effects exist, the concept of equity, however, is much more complicated

to “grasp”.

For example, an equal distribution of costs and benefits suggests an equitable

final effect. But as Harrison and Seidl (1994) point out, fairness does not automatically

correspond to equal distribution because people do not necessarily prefer the “most equal”

result. Instead, additional factors significantly influence the perception of fairness and they

vary among different individuals as well as socioeconomic groups.

Therefore, it becomes

obvious that even if one single concept of equity existed, it would hardly satisfy the

preferences of all social entities involved. Thus, even though various concepts of what

Eliasson and Mattsson (2006) refer in their literature review to studies conducted by Arnott et al. (1994),

Glazer and Niskanen (2000), Richardson (1974) and Small (1983).

Ochmann and Peichl (2006) present four sub-concepts of distributional effects: inequality, polarisation,

progression in taxation and poverty and richness, where they propose common descriptive measures of

inequality like the Gini-coefficient and the Lorenz curve as well as other variance measures. Franklin (2004)

points to quantile-quantile plots and relative distributions and polarization indices as methods used for

inequality and distribution assessment techniques. Abe (1975) assesses distributional changes by calculating

the weighted consumer surplus, and Feldstein and Luft (1973) use linear programming to quantify consumer

surplus. Furthermore, Thurman and Wohlgenant (1989) use a general equilibrium demand curve to calculate

the surplus; single equation error correction models are an appropriate method to calculate short-run changes

(Kelly, 2005).

The contribution of Harrison and Seidl (1994) tries to link distributional and equity aspects. The authors

depict that differing perceptions of equity among the test persons complicate the introduction of axioms as

measures of equity of distributional impacts. Similarly, Atkinson (1970) criticises the current properties of the

social welfare function lacking the proximity to actual social values, which in turn is one of the difficulties

equity research is concerned with.

could be meant by the terms “equity”, “fairness” or “justice” exist, a universal consensus

has not been found. Attempts to determine overall satisfying concepts are based on

qualitative reflections rather than quantitative methods.

Concluding, the discussion about equity concepts and equity assessment shows that its

overall “construct” is very difficult to be captured quantitatively by the economic theory.

Since there does not exist an indisputable consensus about the definition of equity, the

concept (of an equitable distribution) lacks an applicable theoretical basis. It therefore

remains questionable if equity can be measured normatively. Within the model

implementation described in Chapter 5 a measure is implemented that accounts for welfare

changes in terms of economic utility from transport as well as overall consumption. Hence,

using the Hicksian welfare index the so called Hicksian equivalent variation is calculated to

indicate the amount of income necessary to compensate an individual (in the pre-policy

situation) in order to reach equality with the post-policy utility level (Just et al., 2004). The

welfare measure is based on an agent-based utility function and therefore all changes in

transportation and economic conditions affecting individuals are considered. It measures

changes in money metric utility between the pre- and post-policy equilibrium, which is the

amount of money required to bring a household back to the same level of utility as in the

benchmark equilibrium following changes in prices in counterfactual equilibrium.

Nevertheless, measuring changes in economic utility leaves the question about the fairness

of the policy outcome unanswered. However, despite the apparent lack of a single, theory-

based and measurable definition of equity, compiling a variety of concepts gives a sound

framework for a normative evaluation of the equity effects from passenger car road

charging in Germany, leaving the (dilemma of the) definition of the fairness norm to the

evaluator.

Among the qualitative research Adams established an equity theory in 1962 which has been elaborated by

Walster, Walster and Berscheid (1978). This theory views fairness as a basic need that navigates social

behaviour. Rawl’s (1985) concept of equity includes political and moral aspects, and King and Sheffrin

(2002) apply a concept that includes even psychological elements based on prospect theory. Quantitatively,

Bibi and Duclos (2003) for example use the average poverty gap to measure fairness, whereas Butler and

Williams (2002) apply in their work cooperative game theory to approach the idea of equity.

3 Computable general equilibrium (CGE) and travel demand modelling – extension

of the research

The general objective of this doctoral thesis is to investigate equity and welfare

distributional effects from car road charging on the private household sector in the specific

case of Germany. The policy impact assessment covers the entire economy with special

focus on the private households and their welfare situation before and after the introduction

of the measure. To give a funded understanding of the mode of functioning of

environmental and transport policy measures targeting (private) car use, especially on their

socioeconomic implications, the overall model of the German economy is extended by a

more disaggregated private household representation. The methodology used in this work is

a computable general equilibrium framework with integrated private household car travel

and public transport demand, differentiated by combined income and spatial residence

characteristics.

The work reveals the effectiveness of road use charging given an initial state of a single-

country economy, being the only approach of its kind carried out for Germany. Differences

in household budgets, motorisation, demographics and other road user characteristics

govern the response to road charging, and thereby determine the factual scope of expected

impacts. Varying scenario assumptions as to reaction parameters and policy reform designs

provide the basis for policy simulations. The work demonstrates the strong dependence of

welfare and equity effects from road user charging on the unique design of the revenue

reallocation scheme. The application of micro survey data representative for the German

population as well as the use of household category specific behavioural parameters in form

of demand elasticities better approximate the practical real life situation of the policy

implementation environment. Results are based on coherent micro and macro databases,

partially obtained through sophisticated modelling work. They provide new insight to the

integrated assessment modelling research in the field of car road pricing policy

implementation and distributional effect.

4 Model and database

The purpose of this Chapter is to give a general overview on CGE modelling and the

particular work done within the area of applied overall economic impact assessment of road

pricing measures in the transport sector under consideration of welfare distribution and

equity effects and using CGE models. Furthermore, a description is given of the

methodology applied and the data developed to obtain the results of this thesis work, where

the underlying objective is to conduct an integrated impact assessment of economic welfare

for Germany, in particular for different households with regard to distributional and equity

effects after the introduction of car road charging.

4.1 The general computable equilibrium model

The framework of a CGE model is a well suited approach to evaluate the socioeconomic

and environmental impacts of different economic policy instruments. The method allows

the introduction of a comprehensive empirical basis; the setting of the model is coherent as

to linkages between the underlying social accounting matrix (SAM), budget constraints,

macroeconomic agents or accounts, etc. and the rigorous theoretical constraints.

Computable general equilibrium models offer a consistent framework based on neoclassical

economic theory for conducting controlled experiments regarding policy scenarios

concerning the economy as a whole.

The general equilibrium (GE) theory is the basis in economic theory of numeric GE

models. The GE theory goes back to Walras (1874) who firstly constructed a general

economic model. GE models are based on Arrow-Debreu economic theory approach

(Arrow und Debreu (1954) that were the first to prove the existence of a general

equilibrium within a competitive economy based on Walras’ GE model (1874).

In General, CGE models include linkages between all sectors within the economy, while

taking into account restrictions such as the limited endowment with different resources.

These models are closed in a macroeconomic sense by including the equalization of

economic accounts. Moreover, different policy interventions can be analyzed

simultaneously. This is substantial characteristic of the CGE modelling framework as the

overall impact might differ from the sum of all the isolated effects.

Most common economic fields of CGE model application are public finance, taxation

issues, international trade policy questions, evaluation of alternative development

strategies, implications of energy policies, regional questions, or issues of macroeconomic

policy. The ex-ante evaluation of policy decision are of particular interest when they affect

areas concerned with the intersectoral allocation of resources and distribution of income.

This is also the case when changes in transport policy take place with regard to investment

in transport infrastructure. Both come with substantial implications for the government

budget as well as income distributional consequences. When high marginal rates of taxation

exist, raising government revenue and redistribution income is associated with substantial

distortionary and administrative costs. Hence, the evaluation of most transport policies

needs to take this cost into account, especially since they generate fundamental general

equilibrium effects.

In general a computable general equilibrium approach should take an array of aspects into

account starting with the clarification of the policy issue to be addressed. Next, the

methodology for the CGE based policy analysis should be specified, including the

definition of the theoretical framework for use of the methodology for transport policy

analysis, i.e., road charging. This encompasses the description of the concrete

implementation of the measure, the assessment approach of the consequences from the

introduction of road charging, and finally the interpretation of the simulation results.

The theoretical basis of the model may be formalised as a theoretical model with the

theoretical statements being expressed in terms of mathematical relationships. In a fully

parameterised theoretical model the mathematical relationships are specified in terms of

parameterised functions so that the model can be applied to a database, including a set of

estimates of the relevant parameters. The specification of a parameterised theoretical model

is an intermediate step in constructing a CGE model. It generally involves the exact

specification of the subsystems and variables of the model and of the parameterised

functional relationships representing the behavioural assumptions. The procedure used to

find values for the free parameters of functional forms is known as calibration (Mansur and

Whalley, 1984). The consistency check of the deterministic calibration procedure is the

replication of the initial benchmark: the calibrated model must be capable of generating the

base-year (benchmark) equilibrium as a model solution without computational work. The

consistency check of the model and the database is done with reference to the required

theoretical restrictions of a general equilibrium. The reproduction of the base year data as a

solution to benchmark check requires a rigorous consistency of parameters and the database

used in the model implementation. Therefore, the initial stage of model-oriented database

construction involves the transformation of often inconsistent raw data into fully consistent

data set. After the benchmark calibration and before the implementation of an exogenous

shock in terms of a policy intervention, the model is assumed to be in equilibrium. Hence,

the construction and application of the (static) German CGE model included several

working steps: the collection and analysis of the relevant data for the model database, the

harmonization, organisation, and compilation of the data into a SAM (format), and the data

adjustment in order to achieve an initial equilibrium, i.e., overall income must be equal to

overall expenditures, producers' revenues have to be equal to total factor income, etc.

For the model application a model closure has to be specified. The closure of the model

depends on the issue to be analysed and it indicates which variables are to be considered as

exogenous (i.e. specified by the model user, e.g., instrument variables), or endogenous (i.e.

values determined by the values of the exogenous variables and the equations of the

model). The exogenous variables induce the change within the modelling framework and

are therefore often referred to as “shocks”. To compute the model solution the number of

endogenous variables has to be equal to the number of model equations. To qualify the

performance or the “sensitivity” of the model, a “sensitivity analysis” is usually conducted

testing the dependence of the model results on different parameter choices.

The implementation of a policy scenario is often referred to as the introduction of an

exogenous shock. After the introduction of an exogenous shock a new equilibrium is

computed by setting the prices and production quantities for all commodities such that

market demand equals market supply for all inputs and outputs. In other words, after policy

shocks (or simulations) a new counterfactual equilibrium is calibrated within the numeric

empirically based modelling framework including an update of the initial database. The

results between the initial equilibrium of the economy and the equilibrium after the policy

(shock) implementation can be reported in percentage changes or in levels of economic

variables. Thus, through model simulations using a fully specified CGE model that has

been programmed in appropriate software and that contains a corresponding solution

algorithm it is possible to undertake quantitative policy analyses. Results obtained after the

policy shock implementation include changes in output quantities, factor use and prices as

well as indicators such as consumer utility or welfare. Welfare measures are usually based

on the underlying utility functions in the model. For the computation of welfare changes

most often the Compensating and Equivalent Variation measures developed by Hicks

(1939) are used. More in-depth technical description of the characteristics of CGE models

including examples can be found in Shoven and Whalley (1992 and 1984) or Ginsburgh

and Keyzer (1997).

4.2 CGE and travel demand modelling

The intention to introduce a nationwide road charging scheme in the passenger car sector

comes together with a multitude of economic, ecologic as well as social questions,

uncertainties, and concerns that need to be considered often simultaneously.

A consistent

and integrated assessment of economic, environmental, and welfare distribution impacts

from road charging implemented in the passenger travel sector is done more accurately

when the overall interactions within the economy as well as the budgetary situation and

travel choice behaviour of private households are taken into account. Information referring

to private household travel can be represented by activity parameters such as number of

trips, mode choice, trip purpose, distance travelled as well as monetary measurements such

as travel expenditure in absolute terms and as proportion of the household dispensable

In terms of sustainability impact assessment the European Union (EU) suggests a "careful assessment of the

full effects of a policy proposal [that] must include estimates of economic, environmental and social impacts"

(EC, 2001).

income. Given the different sorts of mobility indicators, travel behaviour in general

including reaction to changes in travel costs – e.g., price increase of car use due to an

imposed road charge – depends primarily on sociodemographic and economic attributes

and choice alternatives. This implies that the sociodemographic and economic profile, e.g.,

the disposable income, as well as the residential location characteristic of an individual or a

household determine the effect, which is imposed by travel supply side changes in form of

price rises.

The integration of different household income and residential location

categories within an overall economic framework of a computable general equilibrium

(CGE) model under introduction of a policy reform such as road charging allows the

investigation of general economic and ecological effects together with accompanying

distributional and equity implications. The numerical property of a CGE model allows for a

quantitative assessment of concerning impacts from a policy measure such as road use

charging. Furthermore, the methodological consistency of the CGE approach gives the

advantage of an integrated comparison of quantified effects in the overall sector context of

“environmental quality, economic performance and income distribution" (Böhringer and

Löschel, 2006, pp. 50-51).

Nevertheless, only a small number of one-country CGE models with focus on passenger

travel demand integration exist. Most CGE models are global or multiregional, focussing

on research questions of international or global relevance, e.g. trade, finance or market

competition policies.

Even fever models tackle the area of passenger travel demand.

Within these few models, where passenger travel demand or passenger transport are

considered, most CGE models are somehow limited with respect to – in particular – the

social impact assessment of policies introduced with regard to passenger road travel.

Existing exceptions are described in Mayeres (1998 and 2004), Broecker (2002), Mayeres

For documented studies on passenger travel demand modeling, car purchase, car ownership and car use

modeling see, e.g., Hautzinger (1978), Dargay (2002), BMVBW (2002), Bresson et al. (2004), Kalinowska et

al. (2005), van de Coevering and Schwanen (2005), Giuliano and Dargay (2006), Johansson et al. (2006),

Kalinowska and Kuhfeld (2006), Limtanakool et al. (2006), Naess (2006) for examples of passenger travel

demand modeling, car purchase, car ownership and car use modeling.

For research work concerned with the topic of CGE model implementation see, e.g., Conrad (1999 and

2001), Bergman (1990), Gottfried et al. (1990), Borges (1986), Kehoe and Kehoe (1994), Klepper et al.

(1995), Pereira and Shoven (1988), Robinson (1999), Shoven and Whalley (1984 and 1992), Fehr and

Wiegard (1996), Piggot and Whalley (1985 and 1991), Bhattacharyya (1996), Gunning and Keyzer (1995).

and Proost (2002), Steininger (2002), Munk (2003), Steininger and Friedl (2004), Schaefer

and Jacoby (2005 and 2006). Finally, no CGE model which accounts for passenger travel

demand disaggregated for household income and residential location categories has been

ever applied to Germany. In those CGE models yet applied to German data, demand for

passenger travel is not explicitly included or it does not allow the evaluation of welfare

distribution (equity) effects on a disaggregated household level (Meyer and Ewerhart, 1989;

Broecker, 2001, 2002, 2004; Bach et al., 2001). To account for this shortcoming a CGE

model including passenger travel demand for Germany – the German Road Travel Policy

Model (GRTPM) – has been constructed based on the Austrian version – the Austrian Road

Pricing Model (ARPM) introduced in Friedl and Steininger (2004).

The GRTP model

represents a small open economy of a single country and it accounts for different private

household types according to income and – as an extension to the original structure of the

ARPM – residential location categories. The modelling framework and the underlying

database are described in detail in Chapter 4.

4.2.1 Description of the German Road Travel Policy Model (GRTPM)

The methodological approach used in this dissertation work is based on standard (static)

general equilibrium (GE) theory. The construction of the German CGE model is based on

the Austrian version – the Austrian Road Pricing Model (ARPM) introduced in Steininger

and Friedl (2004). Moreover, a German database was constructed for the standard CGE

model and extended by a disaggregated incorporation of the private household sector. The

constructed model will hereafter be referred to as the German Road Travel Policy Model

(GRTPM).

The underlying model code is written in the Mathematical Programming System

for General Equilibrium Analysis (MPSGE), which is a programming language designed

by Rutherford (1999) in late 80s for solving Arrow-Debreu (Arrow and Debreu, 1954)

economic equilibrium models (Paltsev, 1999).

21 The ARPM has been developed and implemented at the Economics Department of University of Graz, it is

documented in Steininger et al. (2007).

For the MPSGE based model specification the Generalised Algebraic Modelling Software (GAMS) is used

(Brooke et al., 1992). GAMS is an optimization software used in general equilibrium model format by

The GRTP modelling framework consists of separated files containing the model code

including the functional forms and some numerical parameters and the database in value

terms. The value terms of the model variables, e.g., production or endowment values

correspond to a quantity and a price variable. Solving the model for its equilibrium

endogenously updates the values for quantities and prices.

35 production sectors are distinguished in the GRTP model, of which the following are

directly linked to the representation of passenger travel demand: extraction of crude

petroleum and natural gas, transport equipment, distribution, land transport, supporting and

auxiliary transport, finance and insurance, as well as other market services. Agents are

modelled using either a representative microeconomic consumption or production function

(Varian, 1993). For an overview the main model equations and variables are enclosed in the

Appendix 1 and Appendix 2.

A consumer is characterized by a preference ordering over the obtainable goods described

by a utility function and by a budget set that is limited by his income. He is assumed to

choose that bundle of goods in his budget set that maximally satisfies his preferences, i.e.,

his behaviour can be described by utility maximization over his budget set. The GRTP

model distinguishes three agents: private households, the government, and the road pricing

agency.

A (non-transport good) production sector j is assumed to use the optimal production

technology that transforms an input bundle consisting of all goods in the economy and of

the primary production factors into an amount of its output good. Production of non-

passenger-transport goods follows a nested constant elasticity of substitution (CES)

structure, with capital and labour as primary inputs, and intermediate inputs entering in a

Leontief functional form (i.e. with substitution elasticity equal to zero). The producer

chooses an input-output bundle that maximizes his profit, i.e., the producer behaves in a

profit maximising (or cost minimizing way) over the production set defined by his

technology. Equations (1) and (2) describe the gross production X

and the factor aggregate

of sector j:

introducing all required model equilibrium conditions into constraints on nonlinear optimization. It has a

convenient and transparent format as to its syntax, or display, explanatory and diagnostic features.

(1)

(

)

min , 1,...,35

j j j ij ij

X H A X a for j= = ,

where A

and a

denote the Leontieff-input-output-coefficients in sector j,

(2)

( )

(

)

1 1

1 1,...,35

j j

j j j j

j j j j j

H L K for j

σ σ

σ σ σ σ

δ δ

−

− −

= + − = ,

with L

as the labour and K

as the capital input in sector j and σ

as the production elasticity

of substitution between labour and capital in sector j. δ

denotes the constant elasticity of

substitution (CES) distribution parameter in sector j.

Agents are assumed to take the prices of the goods as given. Prices in turn satisfy the

market-clearing criterion, where total demand equals total supply. Thus, the equilibrium

conditions imply that markets for goods, labour and capital clear, firms receive zero exess

profits in equilibrium, and income is equal to expenditure for all households.

The production of the passenger transport intermediate and final consumption goods as well

as household demand are of special interest for this work and are therefore described

extensively in Chapter 4.2.2 as to the functional specification and in Chapters 4.3.3 to 4.3.5

as to the empirical representation.

The government sector is endowed with income from direct (ad valorem on final sales)

taxes and indirect taxes on production, capital and labour (income) as well as from fuel tax

and vehicle tax collection and a social insurance premium contribution from private

households. The government budget is spent on public consumption, investment and

household net transfers of social benefits using a constant expenditure shares form.

Moreover, the government has the function of affecting the distribution of income through

road charging revenue financed household transfers. Government budget balance is a

property of the equilibrium condition.

The foreign trade is subject to the Armington (1969) assumption of product differentiation.

The input cost on each composite good is decomposed into the cost of obtaining the

The current model incorporates all standard equilibrium conditions that are characteristics of the

computable general equilibrium literature (Shoven and Whalley, 1992; Ginsburgh and Keyzer,1997).

domestically produced variant of this good and the cost of the imported variant of this good

using a constant elasticity of substitution function with a finite elasticity of substitution

between the domestically produced variant and the imported variant. Hence, a change in the

price relation between foreign and domestic goods is followed by a trade balance shift

according to the sector specific foreign trade elasticity. Equations (7) and (8) denote foreign

trade in the GRTP model:

(7)

(

)

1,...,35

o w

j j j j

EX EX P P for j

= =

with EX

denoting export of sector j at the quantity EX

for the reference year 0, P

being

the world market price of goods aggregate X in sector j and P

being the (domestic)

production price of goods aggregate X in sector j, and ε

describing the foreign trade price

elasticity parameter of demand in sector j;

(8)

(

)

1,...,35

o w

j j j j

M M P P for j

= = ,

with M

denoting import of sector j at the quantity M

for the reference year 0.

For the analysis of the labour market impact of the transport policy introduced, it is

assumed that the labour market does not clear and the unemployment is determined by

classical, i.e. high minimum wage unemployment (equation (9)):

(9)

≥ ⊥

low

w u

The road charging measure is implemented as a distance dependent markup on the price of

car travel calculated based on the kilometres travelled. The overall effects of the charging

policy depend on the reallocation of the revenues collected from its implementation.

Implemented policy scenarios as well as the road charging revenue redistribution policies

are subject to simulation analyses and are described in detail in Chapters 5.1 and 5.2.

The model is closed by a fixed foreign trade balance at the level of the reference year

("neoclassical closure", fixed foreign savings), such that investment is savings driven and

the foreign exchange rate adjusts to achieve equilibrium.

4.2.2

Private household sector, travel demand and transport production

Private household demand is represented by a nested CES structure, with unity elasticity of

substitution for consumption goods and services other than passenger travel. In equation

(10) C

he,r

denotes the total consumption of household type h

e,r

, differentiated by income

level and/ or residential location attribute. Total household consumption is the sum of X

he,rC

being the consumption of non-transport goods of household h

e,r

and T

he,r

being the transport

consumption of household h

e,r

he,rC

denotes the CES-distribution parameter in non-

transport consumption for household h

e,r

and

he,rC

the elasticity of substitution between

transport and non-transport demand for household h

e,r

(10)

( )

(

)

,,, ,

, , , . .

-1

 

−

 

 

−

 

= 

 

 

+ −

he r

Ce r

C C

he r e r h h

e r e r

e r e r e r e r e r

C C

h h h h h

C T

σ σ

δ δ

11 14

41 44

 

 

= 

 

 

M O M

e r

h h

for h where e is the income category and r the resi

dential location attribute

h h

Consumption of non-transport goods of household h

e,r

follows equation (11) with δ

he,riX

denoting the CES-distribution parameter in non-transport consumption for household h

e,r

and σ

he,rX

the elasticity of substitution between non-transport goods in household h

e,r

consumption:

(11)

(

)

( )

, ,

, , , ,

, , ,

 

−

 

 

−

 

 

= =

 

 

 

 

 

∑ ∑

X X

h h

e r e r

X X

h h

e r e r

e r e r e r e r

c X c X

h h i h i h i

X with

σ σ

δ δ

For reasons of simplification, it is furthermore assumed that all households have the same

consumption structure within the bundle of non-transport goods. This assumption is as far

plausible as most of the consumption aggregates are sufficiently similar across the

household categories defined for the purpose of this study. For the transport goods

aggregate household category specific consumption structures are specified.

he,r

is the Transport consumption of household h

e,r

with δ

he,rT

as the CES-distribution

parameter in transport consumption for household h

e,r

and σ

he,rT

as the elasticity of

substitution between private car transport and public transport demand for household h

e,r

shown in equation (12):

(12)

( )

(

)

, ,

-1 -1

, ,

, , , , ,

   

   

   

−

 

 

= + − +

 

 

T T

h h

e r e r

T T T T

h h

e r h e r h

e r e r

e r e r e r e r e r

T p T u

h h h h h

T T

σ σ σ σ

σ σ

δ δ

The demand between the bundle of non-transport and the bundle of transport goods and

services is governed by the calibrated elasticity of substitution of δ

=0.275, used across all

household categories defied within the model.

A calibrated elasticity of substitution δ

=0.636 is used to express the relationship of the

demand for the two different passenger transport goods, i.e., demand for motorized

individual car travel and for public transport.

Through the specification of elasticity of substitution parameters adapted from studies on

country specific micro-econometric travel demand modelling (BMVBW, 2002), we

consider the quantitative extent of the reaction potential to changes in the price of car travel

induced by road charging.

The use of uniform substitution elasticities across household income groups implies that

An elasticity of substitution between private and public transport with a value of 0.5, for example, means

that a 1 % price rise of car travel relative to public transport induces a 0.5 % change in the modal split, here in

favour of public transport.

For extensive literature surveys of car use and ownership elasticities see also Goodwin (1992), Johansson

and Schipper (1997), Blum et al. (1988), Graham and Glaister (2002a, 2002b and 2004), Goodwin et al.

(2004).

observed distributional effects from the road charging measure are induced by the

household specific travel behaviour in terms of expenditure and activity parameters

(Steininger and Friedl, 2004).

The elasticities of substitution governing the demand between the bundle of non-transport

and the bundle of transport goods and services (δ

=0.275) as well as the bundle between

private and public transportation (δ

=0.636) imply a fairly inelastic demand for everyday

mobility. Hence, despite of the road charging induced price increase of private car travel,

the demand for car kilometers will be only little affected. Therefore, the CGE simulation

effects on the economy and welfare (re)distribution can be expected to be rather strong,

where the extent of welfare redistribution is linked to the “generous” amount of revenues

from the road charging policy that is in turn realisable also because of the fairly price rigid

travel demand.

In case of easier substitution between transport and non-transport goods and/or between car

and public transport, negative impacts on economic welfare as well as on output could

ceteris paribus turn out to be weaker; i.e., the reduction in car travel demand would be

higher due to the road charge measure causing also higher emission cutbacks. Given higher

substitution elasticities for car travel also lower road charging revenues would have to be

expected leading to a reduced redistribution policy effect.

The production of the passenger transport intermediate and final consumption goods

follows the equations (13) to (16). Figure 2 illustrates the structure of household demand.

Figure 2 Structure of household h final demand

Source: Steininger and Friedl (2004).

Household travel demand enters the model in a disaggregated fashion by household

category as to income and/ or residential location characteristic through data on transport

expenditures and on transport activity quantified by passenger and vehicle kilometres

travelled by mode and per year. Passenger transport consists of private and public travel

expenditures and can be considered as being a newly constructed good within the

production and the final consumption of private households.

As said before, identical structure but varying absolute levels are assumed across household

groups with regard to all remaining consumption expenditures. This simplifying

assumption is in line with findings based on data from the German Sample Survey of

Income and Expenditure.

Private travel expenses consist of variable household expenditures on car use and of fixed

household expenditures on car purchase and ownership. The first category depends almost

Source: Einkommens- und Verbrauchsstichprobe (EVS) 2003 (StaBuA, 2005a, 2005b, 2005c, and 2006).

entirely on household specific car use patterns and combines expenditure on car fuels, fuel

taxes and levies, car repair and maintenance costs, and different kinds of costs for parking.

Private household demand for these inputs is satisfied from corresponding (intermediate)

sectors within the input-output table, such as the refined oil, transport equipment,

distribution, finance and insurance, and inland transport.

Hence, all cost components of

household travel demand are linked with the corresponding economic supply sectors of the

German input-output table and social accounting matrix as well as the government budget

in the case of vehicle or gasoline taxes.

Private car travel production is expressed through equation (12) to (14):

(13)

(

)

min ,

pf pf pv pv

A A

T T T= ,

(14)

(

)

min

i i

X A

T= , and

(15)

(

)

min ,

pv p kmp

i i

X A km AT= ,

where T

denotes private car passenger transport with T

being the fixed and T

the

variable, directly kilometre dependent input in the production (and consumption) process of

private car passenger transport. A

, A

pfi

, and A

, A

pvi

are the corresponding Leontief-

input-output-coefficients in private car passenger transport. A

kmp

denotes the kilometre

input coefficient in private car passenger transport and km

are the vehicle kilometres

driven by households in private car transport. Household demand for car travel is satisfied

from the combination of fixed and variable inputs, where the corresponding cost

components follow a Leontief function with an elasticity of substitution set at zero (see

Figure 2). This implies that kilometre charges applying to the variable input cannot be

substituted by other fixed input components, i.e., there are no technical devices to avoid

kilometre charges other than reducing the driving activity.

The German database in form of a social accounting matrix has been constructed based on the input-output

table and other information available from the German Federal Statistical Office (for detailed description see

Chapter 4.3).

The production of public passenger transport T

follows a Leontief structure with A

input -output -coefficient in public transport:

(16)

(

)

min

i i

X A

T= ,

where X

stands for inputs from sector j.

The data used for the representation of the private household sector were mainly derived

from two separate data sources: the German Sample Survey of Income and Expenditure

(Einkommens- und Verbrauchsstichprobe (EVS) 2003, StaBuA, 2005a, 2005b, 2005c, and

2006) and the survey data from Mobility in Germany (MiD, 2002). The exact processing

and combining of the household micro data to generate household categories as to income

and residential location as well as the data sources are described in detail in Chapters 4.3.3

to 4.3.5.

4.3 Database construction

For the construction of German travel patterns for different household income groups,

several data sources were used: the German Sample Survey of Income and Expenditure

(EVS 2003), the Continuous Household Budget Survey (Laufende Wirtschaftsrechnungen

(LWR) 2003, StaBuA, 2005a, 2005b, 2005c, and 2006a), German Input-Output Matrix

based on National Accounts (Volkswirtschaftliche Gesamtrechnungen (VGR) 2000,

StaBuA, 2006b) and finally survey data from Mobility in Germany (MiD, 2002) and The

Car Mileage Survey (Fahrleistungserhebung, 2002, BASt, 2005). In the following the main

data sources, the merging of the data, and finally the construction of different household

income and residential location categories are described in detail.

4.3.1 Input-output data

The core data basis of a CGE model is the social accounting matrix (SAM) of the economy

considered. The fundament of the SAM is the input-output table derived from economic

supply and use tables. The symmetric input-output table with regard to the classifications or

units used in both rows and columns is the result of balanced supply and use tables for a

given point in time, e.g., a base year.

Input-output tables portray in a detailed and clearly laid out way the complex processes of

production and the use of goods and services (products) from domestic production and

imports, the use of goods and services for primary inputs (labour, capital, or land),

intermediate consumption and final use (consumption, gross capital formation, exports),

and the corresponding income generation within an economy. Income or revenue of the

government is obtained by tax and tariff collection. Input-output tables show therefore the

structure of the costs of production and income generation in the production process, the

flow of goods and services produced within the national economy, and the flows of goods

and services with the rest of the world. Components of value added such as compensation

of employees, other net taxes on production as well as consumption of fixed capital, net

operating surplus are also considered within the input-output framework. The format of

symmetric input-output tables can either be made on the basis of an industry by industry or

product by product classification. The classification applied for industries is the General

Industrial Classification of Economic Activities within the European Communities

(NACE). Products are classified according to the Classification of Products by Activity

(CPA). The coding systems of both classifications are compatible with each other. The

selection of the classification type of input-output tables (product by product vs. industry

by industry) depends on the specific objective of economic analysis (Eurostat, 2008).

Figure 3 shows the systematic setup of a symmetric input-output table according to the

product by product classification as has been used to construct the SAM for the GRTPM.

Homogeneous Branches

Products (CPA)

No 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

1 Products of agriculture

2 Products of industry

3 Construction work

4 Trade, hotel, transport services

5 Private services

6 Other services

7 Total at basic prices

8 Direct purchases abroad by residents

9 Purchases on the domestic territory by non-residents

10 Taxes les subsidies on products

11 Total at purchasers’ prices

12 Compensation of employees

13 Other net taxes on production

14 Consumption of fixed capital

15 Operating surplus, net

16 Value added at basic prices

17 Output at basic prices

18 Imports CIF intra EU

19 Imports CIF extra EU

20 Imports CIF

21 Supply at basic prices

Source: Eurostat Manual of Supply, Use and Input-Output Tables.

Final demand at basic prices

Exports Intra EU FOB

Exports extra EU FOB

Total

Total use at purchasers' prices

Final consumption expenditure

by government

Gross fixed capital formation

Changes in valuables

Changes in inventories

Other services

Total

Final consumption expenditure

by households

Final consumption expenditure

by non-profit organisations

Homogeneous Branches Final Uses

Current and constant prices

Intermediate

consumption

at basic prices

Agriculture

Industry

Construction

Trade, hotel, transport

Private services

Imports CIF Empty

Valua added at basic

prices Empty

Figure 3 Symmetric input-output table at basic prices (product by product)

The first quadrant of the table contains numeric data on the intermediate inputs used to

produce each product. The second quadrant contains the data on the final use of products in

the economy. Columns indicate the final demand of a product. Data in the first and second

quadrant are not differentiated by the origin of the inputs, i.e., domestic vs. imports. Trade

and transport margins as well as taxes and subsidies are explicitly accounted for in the

input-output matrix. Therefore, data in the table are valuated at basic prices. Value added is

distributed across the different products according to its origin. Finally, imports are added

by each product category to the domestic output adding up to the total supply in the

economy.

Final uses are outputs of economic activities exiting the economic cycle and divided into

private consumption in terms of the purchases of private consumption goods and services

as well as goods and services from non-profit institutions, government consumption in

terms of all publicly provided service, fixed capital formation including change in stocks in

terms of investment goods employed in the production process, and exports of goods and

services. In general, it is assumed that consumer durables are consumed in one period.

Traditionally, the flows between the national economy and the rest of the world only record

goods and services, excluding, e.g., externalities.

The symmetric monetary input-output framework coherently and systematically links data

of industries, products and sectors. It is balanced when total input equals total output for a

given column and row of products and total value added equals total net final expenditure.

The input-output table provides the most important macroeconomic aggregates such as

gross domestic product, value added, consumption, investment, imports and exports. To

calculate the gross domestic product different approaches can be used in the input-output

framework: the production approach, the income approach, and the expenditure approach

(Eurostat, 2008). A systematic overview of the gross domestic product calculation

approaches is presented in Appendix 3.

The production approach is applied to compute value added and the gross domestic product

for all industries on an annual basis. In general it takes into account the contribution of each

economic unit to production, i.e., the value of their total output less the value of employed

inputs. It is based on information on total output at basic prices, intermediate consumption,

and product taxes less subsidies. Data required for the production approach come from the

annual business surveys, sales, purchases, inventories, gross fixed capital formation,

employment cost, agricultural data and general government non-market data.

Based on the income approach gross domestic product is calculated as the addition of the

different components of value added, i.e., taxes and subsidies on products, compensation of

employees, other net taxes on production and gross operating surplus. The gross domestic

product is the total of all income earned by resident individuals or businesses in the

production of goods and services. Alternative to the income approach is the expenditure

approach. It is used for calculation of the government final consumption expenditure based

on government accounts and for exports and imports of goods and services based on

foreign trade statistics and balance of payments statistics. It also computes the household

final consumption expenditure and gross fixed capital formation. Final consumption

expenditure of private households is estimated from a matrix of the main consumption

activities of private households with products according to the classification of production

activities (CPA) in the rows and the main categories of the classification of individual

consumption by purpose (COICOP) in the columns using data from consumer expenditure

surveys and retail trade statistics. The expenditure approach measures total expenditure on

final goods and services produced in the domestic economy. It is equivalent with the sum of

final uses of goods and services by resident institutional units less the value of imports of

goods and services. The integral procedure of construction of input-output tables from use

and supply data ensures the equality of the GDP estimates using the production, income,

and the expenditure approach (Eurostat, 2008). The flowchart in Appendix 4 gives a

detailed overview of the mechanism and data requirements underlying the construction of

national accounting tables including input-output tables.

A social accounting matrix can be actually understood as an extended input-output table or

national accounting matrix. The term social accounting matrix has a long history going

back to the end of forties and beginning of the fifties (Stone, 1949, 1951, 1955a, and

1955b; Stone and Croft-Murray, 1959) and its definition has been gradually changing. The

social accounting matrix emphasises the role of people in the economy representing them

by a more detailed breakdown of the household sector or a disaggregated representation of

the labour market taking into account various categories of employed persons or

sociodemographic (e.g., age, sex or educational attainment) as well as economic categories

The integration of human activities and interrelations of income and transfer flows between

the different institutional units – among them private households, non-profit institutions

serving households, and government – are other examples of the inclusion of the

socioeconomic aspect into the economic input-output or national account analysis. For the

input-output framework the consideration of the human factor can also mean that the final

use or the final consumption expenditures by product group are further split by institutional

sectors, such as household types, household sub-sectors, or other population groupings

(Eurostat, 2008).

Input-output data used for the construction of the GRTPM social accounting database were

provided by the German Federal Statistical Office (StaBuA, 2004) and corresponded to the

generally promoted product by activity classification that is regarded as being more

homogeneous concerning the description of the transactions. The data are part of the

German national accounting and integrated into the European System of Accounts (ESA

1995), which is fully consistent with the world wide System of National Accounts (SNA

1993).

The square matrix format of the input-output matrix within the social accounting table

allows sector disaggregation according to the application of the model.

The database for the German road travel policy model has a standard formulation of a

Social Accounting Matrix and consists of 35 sectors, as listed in Table 2.

Table 2 Sector data of the GRTPM

Input-output sectors in the GRTP model,

passenger travel demand related sectors are highlighted

No. CPA*

1. 01,02,05 Agriculture and Forestry

2. 10 Mining of Coal and Lignite

3. 11 Extraction of Crude Petroleum and Natural Gas

4. 23 Manufacture of Refined Petroleum Products

5. 40 Electricity, Gas, Steam and Hot Water Supply

6. 41 Collection, Purification and Distribution of Water

7. 27 Ferrous & Non Ferrous Metals

8. 13,14,26 Non-metallic Mineral Products

9. 24 Chemicals

10. 28 Metal Products

11. 29 Agricultural & Industrial Machines

12. 30 Office Machines

13. 31,32 Electrical Goods

14. 34,35 Transport Equipment

15. 15,16 Food and Tobacco

16. 17,18,19 Textiles, Clothing & Footwear

17. 20 Timber & Wood

18. 21 Paper

19. 22 Printing Products

20. 25 Rubber & Plastic Products

21. 37 Recycling

22. 33,36 Other Manufactures

23. 45 Construction

24. 50,51,52 Distribution

25. 55 Hotels and Restaurants

26. 60 Land Transport

27. 61,62 Water and Air Transport

28. 63 Supporting and Auxiliary Transport

29. 64 Communications

30. 65,66,67 Bank. Finance & Insurance

31. 70,71 Real Estate

32. 72 Software & Data Processing

33. 73,74 R&D, Business Services

34. 92,93,95 Other Market Services

35. 75,80,85,90,91 Non-market Services

Source: Database GRTPM, and own calculations.

* CPA (Classification of Products by Activity) is the European statistical classification of products linked to branches of

economic activity in the European Economic Community, Edition 2002.

Corresponding to the schematic presentation shown in Figure 3 additional economic data

such as direct taxes (on labour and capital), primary inputs like labour and capital, imports,

and consumption, were included basically from the German national accounting data of the

Federal Statistical Office to construct the database for the GRTPM.

4.3.2 Transportation sector input-output data

The economic input-output data related to private household travel activity expenditures

were constructed based on the 2003 survey data of income and expenditure of households

in Germany. Data representing private travel demand related household consumption

expenditures were disaggregated according to household income and residential location

(see Chapter 4.3.5) and attributed to the corresponding sectoral production activity of the

GRTP model’s SAM. Both data sources – the income and expenditure survey information

and the input output national accounting data – were integrated for consistency purposes of

the overall SAM database. The detailed illustration of all data incorporated into the model

database as well as the disaggregation of the data is presented in Figure 4.

Figure 4 Social accounting matrix of the GRTPM – attribution of private

household travel expenditures to the sectoral input-output database

The relevant numerical information on differentiated private household expenditure on

travel activities as a consumption activity was derived from the micro data of the German

Sample Survey of Income and Expenditure where it is calculated based on the classification

by consumption purposes and included the indirect value added tax (VAT). It was then

allocated to the sector systematisation of the input-output table by classes of production

activities, excluding indirect taxes (see Figure 4). The correction for the VAT was made

based on the assumption of implicit VAT rates. These were calculated from the difference

between the sector specific private household consumption given in the national accounting

framework compared with the corresponding values obtained from the income and

expenditure survey data, after its reclassification from consumption purpose to production

activity.

To obtain an accurate picture of the household car travel expenditure the following

consumption purposes were taken into account:

Purchase of new vehicles,

Purchase of second-hand vehicles,

Spare parts and accessory for motor vehicles,

Fuels and lubricants,

Maintenance and repair of motor vehicles,

Rental fees for parking and garages,

Other services related to car usage,

Insurance and financial services, and

(Annual) Tax on motor vehicles.

Information on the amount of fuel tax was separated from the household expenditure on

fuels. Fuel as well as motor vehicle taxes were allocated to the government budget revenues

(see Figure 4). Furthermore, expenditure on car travel was grouped into fixed and variable

expenditures. Variable expenditures are directly related to car use, i.e. car kilometres and

comprise expenditures on fuels and fuel taxes. The remaining expenditures make up the

fixed costs of car use, i.e. annualised car purchase and car maintenance costs.

Household expenditure for public transportation was derived from expenditure attributed to

the consumption purpose “external transport services without aviation and without holiday

trips”. It includes expenditures on passenger transport by local and long-distance bus and

coach services, urban light railways and tramways, mountain and funicular railways, taxis

and chauffeur-driven hire cars. Hence, on household demand side separated consumption

expenditures – private and public transportation – were merged into one combined

passenger transport good of final household consumption. Expenditures for aviation and

holiday travel were excluded.

4.3.3 Mobility data

To cover the need for information on household travel behaviour, primarily data from the

National Travel Survey Mobility in Germany (MiD, 2002) were used. To check the

coherence of selected mobility indicators additional data from the Car Mileage Survey

(Fahrleistungserhebung, 2002) were taken into account. This data source will not be

discussed here in detail due to its secondary importance for the GRTP model data base

construction process.

The methodological groundwork and the technical realisation of the MiD 2002 survey was

conducted by the German Institute for Economic Research DIW Berlin and the Institute for

Applied Social Sciences infas, Bonn on behalf of the Federal Ministry of Transport,

Building and Housing.

The MiD 2002 survey was carried out by mail and telephone, where the interviewer

collected information on the daily travel and other activities recorded in a diary for one

randomly assigned day as well as the economic and sociodemographic attributes of the

household members. Sampled households were chosen through a random selection from

population registers of 300 municipalities stratified by types of regions. The register

provides information on names, address, gender, age, and nationality of the survey unit.

The sample contained 25,000 households. Institutionalized persons, foreigners, and

children were included in the sample if they were registered within the municipality. The

sample provides an observed database of 61,700 persons in 25,800 households owning

34,000 cars and reporting 168,000 trips, covering all days of the year.

The MiD is a large-scale, multipurpose cross-section survey financed and supervised by a

national authority. The extensive data contained in the MiD 2002 allow the analysis of

individual and household travel behaviour in the context of household income,

sociodemographic or land use characteristics, accessibility (mixed-mode), or the

From the beginning on the entire project Mobility in Germany MiD 2002 has been documented on its

homepage including detailed information and downloads of interim and final reports (www.mid2002.de). The

survey micro data are available via the Clearing House for Transport Data (www.clearingstelle-verkehr.de).

interactions of travel patterns on the household level, estimation of population totals, e.g. in

car ownership, the amount of travel, vehicle kilometres, mode use, etc. Figure 5 gives an

overview of the information included in the MiD 2002 survey.

Figure 5 Contents of the Mobility in Germany 2002 survey

Source: U. Kunert and Follmer, R. (2005).

4.3.4 Household budget and expenditure data

The Sample Survey of Income and Expenditure (EVS, 2003, StaBuA 2005a, 2005b and

2006a) and the Current Household Budget Survey (LWR, 2003, StaBuA 2005c) are

components of the voluntary system of household budget surveys. Both surveys provide

important official statistics on the standards of living of households in Germany. The

Sample Survey of Income and Expenditure was initiated in 1962/63 and is conducted every

five years in cooperation between the Federal Statistical Office and the statistical offices of

the German Laender. The Federal Statistical Office is responsible for the survey technical

and administrative organisation as well as for the further reprocessing of the data; the

German states recruit the households and conduct the interviews. The current household

budget survey reports results on an annual basis and played a minor role in the GRTP

model database construction. It has basically served as a secondary data source to

crosscheck the consistency of the results obtained from the Sample Survey of Income and

Expenditure and it will not be described here in detail.

For the purpose of representativity the survey covers every five years approximately 60,000

households of all social groups in Germany.

This number includes about 12,000

households in the New Laender and the former East-Berlin. The selection of the survey

population follows a quota sampling plan based on quota variables and applied to the

universe of households. By definition, the Sample Survey of Income and Expenditure does

not provide data on persons living in communal establishments and institutions, neither on

households with a monthly net income of over 18,000 Euro. The reason for putting a

ceiling on the income category are privacy and representativity concerns resulting from the

small number of observed survey units falling into this category. The quota variables such

as type of household, social status of the main income earner, and net household income

determine the number of households to be interviewed. The Sample Survey of Income and

Expenditure comprises three components. The first part of the survey is a postal, reference

day-based introductory interview completed by the households. It captures the basic

sociodemographic and economic data of households and individuals, information about

their housing situation and equipment with consumer durables. The introductory interview

also contains a separate questionnaire about the tangible and financial property, consumer

The legal basis for the sample survey of income and expenditure is the Law on Household Budget Statistics

in the amended version published in the Federal Law Gazette, Part III, Subsection No. 708-6, amended by

Article 10 of the Law of 14 March 1980 (Federal Law Gazette I, p. 294), in conjunction with Article 2 of the

Ordinance of 26 March 1991 (Federal Law Gazette I, p. 846) and the Federal Statistics Law of 22 January

1987 (Federal Law Gazette I, pp. 462, 565), last amended by Article 16 of the Law of 9 June 2005 (Federal

Law Gazette I, p. 1534). The data are collected as specified by Article 2 of the Law on Household Budget

Statistics

(http://www.destatis.de/jetspeed/portal/cms/Sites/destatis/Internet/EN/Navigation/Statistics/Wirtschaftsrechnu

ngenZeitbudgets/WirtschaftsrechnungenZeitbudgets.psml, 21.01.2009).

credit and mortgage debts of households. Both questionnaires are posted to the households

at the beginning of the survey year. Within the second part of the survey households

continuously (diary based) record their income and expenditure over a period of three

months. To ensure an even distribution over the year the sample is subdivided into four

parts, each for a given quarter of the reference year. The third part of the survey is the so-

called "detailed log book" where every fifth participating household enters its detailed

expenditure on food, beverages and tobacco by quantity and price for a given month of the

year. Data on household income and expenditure are recorded according to the international

classification of individual consumption by purpose (COICOP) and is available in the

micro-data as continuous variables in Euro. This allows construction of household

equivalent income quartiles as used for the GRTP model. For the construction of household

income shares different income components such as labour, capital and government

transfers were taken into account. The Sample Survey of Income and Expenditure contains

comprehensive data on the households' different sources of income, their property and debt

situation, equipment with consumer durables, and their final consumption expenditures. It

is therefore regarded as the reference data source for the construction of household

equivalent income quartiles later applied to the mobility data of the Germany Travel

Survey. For an extensive description on the construction of household equivalent income

quartiles see Chapter 4.3.5.

Data on household expenditure contain information on consumption spending as well as

other expenditures, e.g., on insurance or vehicle taxation.

The anonymised and coded data contain also household information on residential location.

Reported land use characteristics are the German federal state (Laender), population-size

categories of municipalities, and a so-called “urban-rural-region” classification, containing

seven classes:

High density agglomeration areas,

Agglomeration areas with outstanding centres,

Urban areas with high population density,

Urban areas with medium population density and a high level centre,

Urban areas with medium population density and without a high level centre,

Rural areas with medium population density, and

Rural areas with low population density.

The “urban-rural-region” classification of households’ residential location is based on the

“basic regional settlement structure typification” (Siedlungsstrukturelle

Regionsgrundtypen) developed and applied by the German Federal Office for Building and

Regional Planning (BBSR, 2005; Schuert et al., 2005). It matches the corresponding

classification applied in the National Travel Survey data and is therefore used to construct

the four residential location classes used in the GRTP model (see Chapter 4.3.5).

In general the “urban-rural-region” classification distinguishes regional patterns based on

population size and population density:

“Basic regional settlement type 1” are agglomerations, characterised by a main centre

area with more than 300,000 inhabitants or by a population density of about 300

inhabitants per km

“Basic regional settlement type 2” are urban areas with a population density over 150

inhabitants per km

or containing a main centre area with more than 100,000 inhabitants

with a minimum population density of 100 inhabitants per km

, and

“Basic regional settlement type 3” are rural areas with a population density above 150

inhabitants per km

but without a main centre area with over 100,000 inhabitants or

containing a main centre area with over 100,000 inhabitants but with a population

density below 100 inhabitants per km

Figure 6 gives an overview of the spatial distribution of the basic regional settlement types

in Germany. 49 % of the German population and 57 % of all employed persons occupy

11 % of the country’s are.

Figure 6 Basic regional settlement structure typification” (Siedlungsstrukturelle

Regionsgrundtypen), German Federal Office for Building and Regional

Planning (BBSR)

To obtain the four residential location classes used in the GRTP model (see Chapter 4.3.5),

parameter values of the “urban-rural-region” variable were reassigned and partly merged:

Categories 1) and 3) were combined to one category “Agglomerations”,

Categories 4) and 5) were merged into the category “Urban areas”,

Categories 6) and 7) were summed up into the category “Rural areas”, and

The residential location category 2) “Urban centres” constituted a unique category.

The residential location attribute used in this work describes the accessibility within the

area occupied by the specified household income category. It has a number of advantages.

Firstly, they are comprised in the two key household data sources: the German Sample

Survey of Income and Expenditure and the Mobility in Germany Survey.

Secondly, the construction of the regional typification used by the German Federal Office

for Building and Regional Planning takes into account different administrative spatial units

such as regions, districts, and municipalities. It is further based on the spatial distribution of

population in terms of the population density and population size as well as the proximity

to and the existence of regional kernels with their centre functions. Both attributes refer to

accessibility potentials of residential areas. Densely populated areas with functional centres

providing residential infrastructure such as social and other services and working activity

areas are assumed to be well accessible, i.e. city centres, urban areas. On the other hand,

areas with low population densities, located relatively far from functional centres indicate

lower accessibility, i.e., rural areas, parts of agglomerations. Accessibility potential

revealed by land use attributes implies the existence of travel mode alternatives as well as

their service quality regarding public transportation. Thus, rural areas are assumed to be

less accessible, especially by public transportation. They are not in direct proximity to

functional centres and can be best travelled by private transport, i.e., car, bike or walking

(see also Appendix 10).

4.3.5 Combining different data sources – private household income, expenditures, and

mobility

To allow the application of the CGE model for evaluating the (re)distributional effects from

road charging policy implementation in the passenger road travel sector, the model

distinguishes among 4 different household income classes, i.e., equivalent income quartiles

and 4 different residential location attributes describing the households. Each household

category is characterized by a uniquely parameterized utility function, endowments of

primary factors such as capital and labour as well as public income transfers and

unemployment benefits. Household primary factor endowment determines its wage and

capital income. For the construction of the model travel demand database by different

household income and residential location categories, travel expenditure and travel activity

data were required. Since no database exists containing both information, two different data

sources had to be used. Hence, household category specific travel demand patterns were

included into the model through behaviour based mobility parameters (in km per mode)

from the Mobility in Germany National Travel Survey. Travel expenditure coefficients (in

€) were derived from the German Sample Survey of Income and Expenditure. Even though

household model data were derived from two different data sources it had to allow for the

assignment into the same income and residential location household population groups.

Furthermore, household equivalent income quartiles had to be calculated for the assessment

of household redistributional and equity effects. The construction of household equivalent

income quartiles requires reliable continuous household income data accompanied by

further sociodemographic household information such as household size and composition

together with the age structure of single household members. This data are required to

derive the equivalent scale that is used to modify the original household net income to

obtain household equivalent income.

4.3.5.1 Construction of equivalence-weighted income quartiles

While data from the Income and Expenditure Survey contain continuous information on

household income and wealth status, income in the National Mobility Survey is reported as

a categorical variable (differentiated for 8 monthly net household income categories,

see Table 3) based on the self-assignment of surveyed households. Thus the German travel

database contains only household income classes and not continuous household income

information. Table 3 gives relevant details underlying the differences in the way household

income is surveyed in the Sample Survey of Income and Expenditure compared to the

National Travel Survey Mobility in Germany.

Table 3 Information available on household income comparing the Household

Income and Expenditure Survey (EVS, 2003) and the German National

Travel Survey (MiD, 2002)

Sample Survey of Income and Expenditure of the German

Federal Statistical Office (Einkommens- und

Verbrauchsstichprobe, EVS, 2003, StaBuA)

German National Travel Survey Mobility in Germany

(MiD, 2002, infas and DIW Berlin)

Precise income is reported during a 3 month period by

sampled household – widely differentiated income categories

are reported, including different types of public and private

transfers as well as lump sum annual bonuses

Surveyed households are asked to assign themselves to

8 income categories of monthly household net income –

voluntary declaration based on self-assessment:

Less than 500 € per month

500 € to less than 900 €

900 € to less than 1,500 €

1,500 € to less than 2,000 €

2,000 € to less than 2,600 €

2,600 € to less than 3,000 €

3,000 € to less than 3,600 €

3,600 € and above per month

Monthly household net income or disposable income are

calculated based on detailed household income structure

reporting

The highest category of “3,600 € and above” of monthly

household net income is not restricted to a maximum

value

Household income is a continuous variable and has unique

observations for each (household) survey unit

The share of households without income reported is

relatively high (14 % of the sample) – this is mainly due to

refused income indication (about 9 %)

In line with the survey design and for the reason of a limited

number of observations households with a monthly net income

over 18,000 Euro are eliminated from the sample

There are no missing values within the income variables –

each survey unit carries income information

Sources: MiD 2002 (infas and DIW Berlin, 2002), EVS 2003 (StaBuA, 2005).

To carry out the analysis of household welfare distributional and equity effects, population

segments (and their consumption and mobility profiles) had to be classified according to

quartiles of equivalence-weighted household income. Hence, the income unit used here is

the household consisting of individuals living together and participating in common

resources, e.g., combined household incomes. The calculation of household equivalent

income takes into account household size and household composition as to the age structure

of the household members. It is based on the idea that household income is earned by

individual household members and after it has been pooled its allocation depends on the

size and structure of the household as a whole. In general, consumption and income

allocation decisions take place depending on the household demographics. Hence,

household income is shared when expenditure and consumption choices are made by

individuals living together in a household context. The implementation of equivalence

scales derived from household size and structure can be also understood based on cost of

living functions derived from economic utility theory, applied to economies of scale for

persons living together (Y=C(U), where income Y corresponds to the cost of living and

therefore the utility or welfare U of the household). For some expenditure purposes and

consumption goods or services household size and its composition as to the age of its

members generate economies of scale with regard to household utility. Several types of

economies of scale can be observed. The concept applies in particular to travel decisions

and travel expenditures. The consideration of household economies of scale through the

calculation of household equivalent income quartiles for the distributional analysis within

the passenger travel sector is therefore essential. From the consumption theoretical point of

view travel choices are made within the household context considering the utility

maximisation of the household as an economic unit. Probably the best practical example is

trip chaining of parents bringing their children to school on their way to work. Therefore,

for the investigation of equity impacts it is essential to take into account equivalent income

to provide a comparison of levels of individual well-being or utility across households, in

particular when travel demand is considered. For a fundamental discussion of household

composition and welfare comparison see also Muellbauer (1973 and 1974).

Equivalence scales are applied to derive the level of income of an equivalent adult, the

equivalent income being weighted by the equivalence scale applied. Hence, the relevant

population unit used in welfare analysis is the equivalent adult. The analytical framework is

therefore based on a derived income distribution of equivalent adults where differences in

(consumption) needs are taken into account corresponding to household formation.

To calculate equivalence-weighted household income the OECD equivalence scale was

used (Atkinson et al., 1995; OECD, 1982 and 2008). The OECD equivalence scale is based

on different weighting factors for adults and children: the main or first household member

receives the factor 1.0, further household members older than 14 years are weighted with

the factor 0.5; finally children at the age of 14 years and younger receive the factor 0.3.

In the next step, income quartiles were computed based on the equivalence-weighted

household income. The extrapolated household survey population was divided into four

equal-sized groups under consideration of the household equivalent income distribution.

Each quartile contains one fourth of the sampled population. Equivalence-weighted

household income quartiles had to be calculated in both household data sources used in this

work, i.e., the Sample Survey of Income and Expenditure and the National Travel Survey,

the latter originally containing only a categorised household income characteristic (see

Table 3).

4.3.5.2 From categorical to continuous income data in the National Travel Survey

Nevertheless, to compute equivalence weighted household income quartiles a continuous

income variable is required together with additional household characteristics. Continuous

household income values were only available in the Sample Survey of Income and

Expenditure. Therefore, this data basis was used as the reference data source on household

income information for the computation of equivalent income quartiles. Furthermore,

explanatory coefficients from the estimation of a household equivalent income regression

function were used to predict the continuous income for the households in the National

Travel Survey data.

For the estimation of a linear regression model of household income a model was specified

with the log-transformed form of the monthly household net income as dependent variable

and significant household characteristics as explanatory variables. In the regression only

comparable explanatory variables enclosed in both surveys could be considered.

Explanatory variables used were: the size of the household/ number of household members,

number of working persons within the household, number of children at the age of 14 years

For example, a family consisting of five members, i.e., two adult parents and three children at the age of 14,

9 and 3 years, has a combined disposable income of overall 5,000 Euro per month. Their monthly net OECD-

equivalence-weighted income amounts to 5,000/(1 + 0.5 + 0.3 + 0.3 + 0.3) = 2,083 Euro.

or younger, number of cars within the household, the German federal state (Bundesland),

and a residential location characteristic as the “urban-rural-region” classification described

in Chapter 4.3.4. The estimation was carried out using the econometric software Stata.

Since all of the explanatory variables are in categorical form a desmat-linear regression

model was run in Stata.

Given the limitations regarding the inclusion of explanatory variables into account, the

model fit can be seen as satisfactory (Pseudo-R-Square of 0.5).

In the next step, data from the National Travel Survey were merged to the Sample Survey

of Income and Expenditure that has been used to estimate the household income regression

model. Since the National Travel Survey data contained exactly the same household

income explanatory variables as have been used in the regression the income predictor was

calculated for both databases using the parameter coefficients estimated from the

Household Sample Survey of Income and Expenditure.

(OECD) Equivalence scales were calculated separately for each dataset using the specific

household demographic characteristics. Based on the estimated household monthly net

income and the calculated equivalence scale parameters, equivalence-weighted household

incomes were calculated for each household in the two datasets – the Sample Survey of

Income and Expenditure and the National Travel Survey.

Finally, based on the equivalence-weighted household incomes, household quartiles were

constructed. Quartiles rather than quintiles or deciles were chosen for the model database to

limit the number of household categories after the inclusion of the four residential

characteristics. The construction of two-dimensional household categories based on, e.g.,

income deciles and the four residential location attributes would lead to 40 different

household categories with a likely critical number of observations in each category.

Besides, the interpretation of distributional effects between 40 two-dimensional categories

could easily become fuzzy and unsound.

Data Analysis and Statistical Software STATA, www.stata.com (24.11.2009).

For a detailed description of categorical regression estimation procedures in Stata using desmat see

http://ideas.uqam.ca/ideas/data/bocbocode.html. (15.07.2009).

4.3.5.3 Matching different data sources

An important precondition for merging information on travel expenditure and travel

behaviour derived from two different data sources for the estimated equivalence-weighted

household incomes quartiles is a good match of the household population within the

quartiles. Therefore, households in the quartiles of each database should show similarities

as to average household size, motorisation levels, age structure of household members, etc.

The method used to estimate household income allows a satisfactory match of the

household population of each survey sample according to the equivalence-weighted

household income quartiles. Small differences between the quartile income intercepts in

each dataset validate furthermore the comparability of the household population

distribution within the household quartiles of both data sources (Table 4).

Table 4 Equivalence-weighted household income quartile intercepts calculated

from the Sample Survey of Income and Expenditure and the National

Travel Survey

Quartile income intercepts, in Euro

Equivalence-weighted household

income quartiles Sample Survey of Income and

Expenditure German National Travel Survey

Quartile 1 up to 1,569 1,496

Quartile 2 up to 1,893 1,840

Quartile 3 up to 2,233 2,141

Quartile 4 from 2,233 2,141

Source: MiD 2002 (infas and DIW Berlin, 2002), EVS 2003 (StaBuA, 2005), and own calculations.

The analysis of distributional and equity effects from road charging will consider not only

household income quartiles, but is also supposed to take into account the four residential

location characteristics as described in Chapter 4.3.4. Therefore, the constructed

equivalence-weighted household income quartiles are additionally combined with their

corresponding parameter values from the “urban-rural-region” classification in each

dataset. For the analysis of the model results categories are distinguished as presented in

Table 5.

Table 5 Household categories used in the application of the GRTPM

Household categories used in the GRTP model

Income

HIQ1 Equivalence-weighted household income quartile 1

HIQ2 Equivalence-weighted household income quartile 2

HIQ3 Equivalence-weighted household income quartile 3

HIQ4 Equivalence-weighted household income quartile 4

Residential location

HRCent1 Urban Center

HRAgglo2 Agglomerations

HRUrb3 Urban areas

HRRul4 Rural areas

Combination of the equivalence-

weighted household income quartiles and the residential location characteristic based on the

“urban-rural-region” classification

H1 HRCent1 + HIQ1

H2 HRCent1 + HIQ2

H3 HRCent1 + HIQ3

H4 HRCent1 + HIQ4

H5 HRAgglo2 + HIQ1

H6 HRAgglo2 + HIQ2

H7 HRAgglo2 + HIQ3

H8 HRAgglo2 + HIQ4

H9 HRUrb3 + HIQ1

H10 HRUrb3 + HIQ2

H11 HRUrb3 + HIQ3

H12 HRUrb3 + HIQ4

H13 HRRul4 + HIQ1

H14 HRRul4 + HIQ2

H15 HRRul4 + HIQ3

H16 HRRul4 + HIQ4

Sources: GRTPM, and own calculations.

All categories used for the application of the GRTP model, i.e., 4 equivalence-weighted

household income quartiles, 4 residential location classes, and the 16 combinations between

income and land use attribute, are consistent between the two datasets as to the underlying

household population. Table 6 shows selected statistics of the sample as to the number of

observations as well as the extrapolated sample together with relevant indicators such as

average household size or income.

Table 6 Selected statistics from the Survey of Income and Expenditure and the

National Travel Survey

Selected household statistics, Germany 2002 and 2003

Number of observations Aggregated total in

million.

Average monthly net

income in Euro

Average number of

persons per household

MiD EVS MiD EVS MiD EVS MiD EVS

Income category

HIQ1 5,457 7,900 9.4 9.8 1,271 1,248 1.7 1.6

HIQ2 6,572 10,423 9.5 9.3 1,753 1,695 2.2 2.3

HIQ3 6,477 11,587 9.5 9.5 2,069 1,983 2.3 2.2

HIQ4 7,342 12,834 9.2 9.5 2,603 2,510 2.4 2.3

Total 25,848 42,744 37.7 38.1 1.920 1.856 2.2 2.1

Residential location

HRCent1 6,757 8,987 9.1 9.2 1,865 1,784 2.0 1.9

HRAgglo2 10,049 18,982 16.5 16.5 2,020 1,947 2.2 2.1

HRUrb3 6,059 9,127 7.4 7.8 1,852 1,812 2.3 2.2

HRRul4 2,947 5,648 4.7 4.6 1,780 1,750 2.3 2.2

Income category together with residential location

H1 1,688 2,289 2.9 3.1 1,281 1,256 1.6 1.6

H2 1,688 1,914 2.3 2.0 1,756 1,686 2.1 2.1

H3 1,508 2,202 1.8 1.9 2,078 1,962 2.3 2.2

H4 1,909 2,582 2.2 2.2 2,586 2,468 2.2 2.1

H5 1,359 2,419 3.1 3.2 1,292 1,269 1.6 1.6

H6 2,340 3,867 4.0 3.4 1,772 1,703 2.2 2.3

H7 2,703 5,759 4.6 4.8 2,079 1,982 2.2 2.1

H8 3,647 6,937 4.9 5.0 2,620 2,524 2.4 2.4

H9 1,617 1,908 1.9 2.1 1,240 1,230 1.8 1.8

H10 1,691 2,781 2.0 2.4 1,722 1,701 2.3 2.3

H11 1,545 2,283 2.0 1.8 2,053 2,006 2.4 2.4

H12 1,206 2,155 1.4 1.6 2,598 2,530 2.6 2.6

H13 793 1,284 1.5 1.3 1,248 1,203 1.9 1.8

H14 853 1,861 1.3 1.5 1,736 1,678 2.5 2.4

H15 721 1,343 1.1 1.0 2,036 1,988 2.4 2.3

H16 580 1,160 0.8 0.8 2,558 2,504 2.5 2.5

Sources: MiD 2002 (infas and DIW Berlin, 2002), EVS 2003 (StaBuA, 2005), and own calculations.

Based on the good comparability of the data within the Sample Survey of Income and

Expenditure and the National Travel Survey it can be assumed that each category defined

within the two survey samples comprises (nearly) the same household population as to

demographic characteristics. Given consistent equivalence-weighted household income

quartiles, including the residential characteristic input data on household income

distribution, travel expenditures, mobility parameters and CO

emissions have been

deducted from the two surveys and integrated into the GRTP model.

Household data from the Survey of Income and Expenditure were used to derive the

income shares for sources of income according to the household categories specified in the

GRTP model as shown in Table 5. Income types taken into account were: overall

household net income, household labour income, income from capital, public transfers and

finally unemployment benefits.

Furthermore, for each household category expenditures on travel related goods and services

including fuel and vehicle tax as well as car insurance were derived from the Survey of

Income and Expenditure database (for a detailed description see Chapter 4.3.2).

From the National Travel Survey mobility parameters were included in the GRTP model

according to household categories. Mobility information for annual kilometres travelled by

car and by public transportation means were used. Moreover, household category specific

emission parameters were computed, based on corresponding vehicle ownership

attributes as reported in the survey.

After the integration of the household data in the model, numerous data checks and iterative

(calibration) adjustments followed to encompass data consistency between the macro, top-

down economic data from the input-output tables of the national accounts and the micro,

bottom-up data from the two household surveys. This is not a trivial exercise, as empirical

data sources often reveal inconsistencies and different sorts of deficiencies.

The GAMS code of the GRTP can be viewed in the Appendix 11.

5 Model implementation

5.1 Policy scenario definition

The implementation of road charging policies requires the specification of the level and the

type of charging. The road use charge is imposed on a distance dependent basis (per km)

and the level of per km charging fee is set based on marginal social or marginal average

road use cost calculations. In the past growing research activity has been dedicated to the

assessment of the full social costs of motor vehicle use, including external, or non-market

costs imposed e.g. on the environment and the private, or market costs born by the car user.

Results from social-costs assessment provide data for the specification of transport pricing

policy measures (Lee, 1993; Murphy and Delucchi, 1998, Delucchi, 2000; Litman, 2003;

Quinet, 2004). Nevertheless, methods used for (full) road transport cost assessment vary

and resulting numerical estimates are often based on different assumptions as to the kind of

costs included into the calculation.

The scenario analysis is carried out based on 0.05 Euro per km, distance dependent road

charge imposed as a mark-up on the variable car travel costs on private car drivers. The

0.05 Euro per km road charge rate is a lower bound, averaged estimate drawn from a survey

of German as well as European studies on road infrastructure cost assessment as well as

external average and marginal social cost calculations. The rate of 0.05 Euro is set as one

half of the reference value from average external cost calculations for cars and is assumed

to correspond to the lower bound from marginal external cost calculations (Herry and

Sedlacek, 2003; IVT, 2004; Infras/ IWW, 2000, 2004; RECORDIT, UNITE). The decision

to investigate a 0.05 Euro per km road charge, rather than a gas tax or a vehicle-specific

charge is clearly related to the given probability of implementation of such an instrument in

Germany, but also to the fact that energy tax on gasoline is already one of the highest

compared to other countries in the European Union, and in particular compared to

Germany’s border countries. A further increase of the tax would worsen the already

Cost accounted for in the calculation can be associated with road congestion, traffic accidents, local as well

as global air pollution, oil dependency, and noise. Other external costs taken often into account are road

infrastructure maintenance costs, land use such as urban sprawl and parking, etc. (Parry et al. 2007). For

detailed description of road use externalities see Chapter 2.1.

existing adverse effect of “grey fuel imports and refuelling tourism”. As the study is future

oriented it focuses on road charging as a policy instrument. At the same time it takes into

account two important policy changes already or soon taking place in Germany: Firstly, the

recent developments within the German vehicle taxation scheme with regard to the

consideration of CO

emissions; and secondly, the European Commission regulation to

come in 2012 forcing European car makers to curb down the CO

emissions of the newly

vehicle registrations to 120 CO

g/km. Both policy regulations will improve fuel

efficiencies of passenger cars, inducing decreasing fuel demand. It can be therefore

expected that in the long run both policy changes will relax the impact of fuel taxation as

policy instrument to regulate car use and generating state revenue and making it therefore

necessary to set up an alternative policy instrument such as road use charging.

The level of 0.05 Euro is additionally linked to the fact that it makes up about 30 % of the

per km cost of car use compared to the out-of-pocket cost of an average car driver in

Germany. It is therefore high enough to trigger behavioural reactions of car users; at the

same time 0.05 Euro per km is a rather small amount compared to the overall household

spending on car purchase, ownership and use. Nevertheless, the implementation of full

social cost pricing would require very high charging levels and therefore is bound to be

controversial.

For the implementation of the GRTPM the following revenue use was assumed. Road

charging is collected and redistributed within the CGE model structure, where 15 % of the

revenues total is retained for system-financing purposes and redirected to intermediary

input sectors such as insurance and banking, electronic devices and the factor labour.

The remaining revenue is divided between a direct transfer to private households and the

investment in public transportation services. The different revenue use scenarios as to

household refund are specified in detail in Chapter 5.2.

Besides the cost associated with operating the road charging mechanisms no additional

costs from transactions or interactions between existing taxes and the introduced

transportation policy are considered. The consideration of optimal tax structures and the

discussion over the revenue share used to cover the operation of the road charging agency

are both complex topics and their in-depth analysis goes beyond the scope of this work. The

assumption about the transaction costs retained to finance the system operating the road

charging revenue collection and redistribution agency has been ever since subject to critical

discussion, in particular when the cost is set as share of the overall revenue. In this study

the approximate value or share of the system cost has been derived from existing road

infrastructure cost assessment studies, given the charge level of 0.05 Euro per kilometer

and corresponding revenues (Herry and Sedlacek, 2003; IVT, 2004; Infras/ IWW, 2000,

2004; RECORDIT, UNITE). The question of an optimal or efficient revenue share of the

road charging agency, especially when simulating changing road pricing levels is not

further elaborated in this dissertation work. It certainly remains a relevant and interesting

research topic.

The follow-up simulation analyses of different road charging revenue recycling schemes

allow the assessment of welfare and equity impacts from such policies, where the choices

of revenue redistribution reflect different policy objectives. The disaggregation of the

private travel demand within a CGE model framework between different transport modes

and household income and residential location categories enables this assessment, where

the final effects from the pricing measure within the economy depend on the use or

reallocation of the monetary returns collected from the road charge (Small, 1992; Meyers,

2000 and 2001; Mayeres and Proost, 2002; Farrell and Saleh, 2005; Hau 1998, 2005a, and

2005b).

The answer to the question about acceptance of the measure strongly depends on the

definition of the acceptability criterion as well as the revenue recycling policy. The GRTP

model allows the simulation of policy scenario assumptions required to design an

acceptable policy reform, provided an applicable fairness and equity definition is assumed.

This includes also the examination of the question how much redistribution has to take

place before the median voter supports a 0.05 Euro per km road tax. At the same time the

methodological approach used has the strong advantage of allowing the assessment of

distributional as well as overall economic and environmental effects important for social

welfare analysis under consideration of revenue redistribution decisions.

5.2 Revenue reallocation schemes – household refund

To examine the differences in the distributional impacts, the overall net welfare, and other

economic effects depending on the road charging policy scenario, the varying road charge

collection and reallocation schemes are specified in the CGE model for Germany. In each

scenario the road charging measure is implemented as a distance dependent mark-up on the

price of car travel calculated based on the kilometres travelled. The overall effects of the

charging policy depends basically on the reallocation of the revenues collected from its

implementation and is therefore subject to policy simulations described below. In each

scenario road charging is collected and redistributed within the CGE model structure,

where 15 % of the revenues total are retained for system-financing purposes and redirected

to intermediary input sectors such as insurance and banking, electronic devices and the

factor labour. 50 % of the revenue flows to the transport sector and is evenly used to

improve or expand the public transport system and the road infrastructure. It is assumed

that the value added for a household group by transport system amelioration financed from

road charging revenue is proportional to the amount of public transport trip-km travelled by

the household. Another possible interpretation of this public transport service and

infrastructure improvement is that it indirectly lowers public transport fares. Depending on

the policy design, the remaining proportion of the road charge revenue of 35 % is

redistributed to the private household sector according to a specific household refund

scheme or remains part of the public household budget.

The different road charging revenue use policies specified to carry out the simulation

analyses with the GRTP are as follows: A all of the 35 % of the revenue is refunded in a

lump-sum manner to the private household sector and distributed evenly, i.e., in equal

proportions across the household categories, labelled “Equal household refund”;

scenario B “Proportional household refund” the private household refund is reallocated to

A lump-sum distribution of the revenue might seem far from the practical implementation being rather

theoretical. Nevertheless, there exist examples of a credit based pricing scheme similar to the lump-sum

revenue distribution scheme, proposed to improve public acceptance of the measure (Kalmanje and

Kockelman, 2004).

100

the household categories according to the specific fuel tax burden of each group as

proportion of the overall fuel tax burden contributed by the private households to the state

budget and therefore to some extent reflecting the “vertical equity” principle often referred

to in public economics (see Chapter 2.3); and finally in scenario C no lump-sum road

charging revenue redistribution to the private households takes place, labelled “No

household refund”. Table 7 gives a summary overview of the three policy scenarios

specified in the GRTPM for the conduction of the policy simulations.

Table 7 Overview of the policy scenarios implemented in the GRTPM

Overview of policy scenarios implemented in the GRTPM

Scenario label A "Equal household

refund“ B “Proportional household refund“ C "No household refund"

Network coverage Full network

Time differentiation None

Charging level 5 Euro-Cent/km

Revenue use 15 % System-financing, 50 % Transport sector, 35 % Private household refund

Household refund

policy

Evenly lump sum

redistribution to all private

household categories

%-Redistribution according to the

specific fuel tax burden of each

household category as proportion of

the overall fuel tax burden

No redistribution to the private

household sector

Source: GRTM.

The redistribution structure of road charging revenues between household categories given

in Table 7 and illustrated in the following Chapter 5.3 does not result from welfare

optimizing assumptions. The simulation of different redistribution schemes of the road

charging revenue is rather used to provide a better understanding of the welfare and equity

impacts from the road charging measure.

101

5.3 Model results

For a comparative static impact assessment a distance dependent and time invariant overall

road network charge for car use is implemented at the level of 0.05 Euro per km. The

implementation of road charging changes the price in car travel and generates a shift in the

modal split resulting in changing overall transport volumes, depending on the reaction

parameters introduced in the model. Furthermore, economic, budgetary and environmental

sectors related to the demand for passenger travel experience an impact from the shifts in

the private travel mode choice.

The effects of road charging can be categorised as follows: 1) higher travel costs for car

users according to the distance driven (in km) on the public road network, 2) rising unite

cost of car use triggers travel behaviour reactions – reduction of car use and changes in

modal split towards “slow travel modes” to avoid the charges, 3) revenue collection and

redistribution to e.g., provision of road infrastructure, public transport service enlargement

or improvments, tax cuts, or public sector spending policy in general. Since the road

charging measure introduced in this work applies to the overall network and does not take

into account the travel time of the day, travel behaviour changes will not include route

choice effects, or changes in departure times. The combination of the three components, in

particular the revenue use policy determine the net effect of the charging scheme and

whether parts of the population will suffer or benefit from the measure (Small, 1992;

Meyers, 2000 and 2001; Mayeres and Proost, 2002; Farrell and Saleh, 2005; Hau 1998,

2005a, 2005b).

5.3.1 Overall transport and macroeconomic impacts

In all road charging scenarios the impacts on the economic activity are rather small. It

suggests that the policy can improve emission levels and shift mode choice towards public

transportation without having negative secondary economic effects of significant

magnitude. Table 8 summarizes the overall transport, environmental and macroeconomic

effects for varying road charge revenue reallocation policies.

102

Table 8 Macroeconomic effects from different road charging scenarios,

Germany 2002

Overall effects from different road charging schemes for Germany

Reference

Equal

household

refund

Change in

Proportional

household

refund

Change in %

No household

refund

Change in

Level of road charge

Euro/ km

0.05

Transport variables

Revenues total

Million Euro

23,042

23,051

22,954

Rev. (semi-

public)

Million Euro

13,075

13,062

19,511

Car

Million km 492,783

460,848

-6.5

461,012

-6.5

459,072

-6.8

Public transport

Million km 133,144

140,725

5.7

140,633

5.6

140,080

5.2

Overall travel

Million km 625,927

601,572

-3,9

601,645

-3,9

599,152

-4,3

Environment

1,000 t 110,698

104,017

-6.0

104,051

-6.0

103,614

-6.4

difference

1,000 t -

-6,681

-6,647

-7,084

Macroeconomic variables

Environmental w

elfare

change

Million Euro

2,568

2,555

2,711

GDP

Billion Euro 2,143

2,171

1.3

2,171

1,3

2,173

1.4

Number of employees 1,000 39,096

39,308

0,5

39,308

0.5

39,384

0,7

Number of unemployed 1,000 4,061

3,849

-5.2

3,849

-5.2

3,773

-7.0

Unemployment rate

9.41

8.9

8.7

Price of capital

-0.04

-0.05

Budgetary effects

Due to change in

Rev. from direct taxe

Million Euro

722,674

725,302

0.4

725,312

0.4

726,284

0.5

Rev. from indirect taxes

Million Euro

60,643

60,857

0.4

60,857

0.4

60,957

0.5

Labour market expend.

Million Euro

43,710

41,432

-5.2

41,426

-5.2

40,615

-7.1

Government demand

Million Euro

378,537

381,071

0.7

381,096

0.7

382,888

1.2

Sources: GRTPM, and own calculations.

The volume of car road charging revenue generated in Germany from charging car drivers

using the overall road network is to a great extent determined by the population size, the

103

number of car users or degree of motorization, the total car kilometre driven and finally the

level of the per km road charge. For the different scenarios there are only slight differences

in the total revenue of about 23 billion Euro, but sizeable differences between 13,062 and

19,511 million Euro, when the semi-public revenue volume is considered. The semi-public

revenue is calculated as the difference between the revenue total and the amounts of

revenues transferred to the road charging collecting agency and the household refund (as

described in Chapter 5.2).

To have a better understanding of the total revenue volume from road charging (23 billion

Euro) it can be compared to the annual volume of car tax or fuel tax collected in Germany.

Hence, in 2002 the German state budget received 7,592 million Euro of car tax and

42,192 million Euro of fuel tax. The approximated road charging revenue would therefore

vary between approximately twice the volume of the annual car tax and be close to half the

amount collected from fuel tax. According to the National Accounts, household

expenditure on overall transport amounted to 165,420 million Euro and for car fuel to

40,380 million Euro for the year 2002. Hence, on the aggregated basis households would

have to pay for road use about one half of their annual fuel expenditure.

Before the implementation of the policy annual car kilometres amount to about 493 billion

km. After the introduction of car road charge a reduction in car use accompanied by an

increase in the use of public transportation can be observed for each scenario. Depending

on the revenue reallocation scheme, private car travel declines between 6.5 % and 6.8 %.

The reduction in car use after the introduction of car road charging is slightly higher, when

there is no direct revenue transfer to the private household sector. The redistribution of a

revenue share directly to the private households lowers the negative effect of road charging

on car use. The decline in auto mobility due to the distance dependent cost rise in car travel

is (partially) compensated by the use of public transit. Kilometres travelled with public

transport modes rise on average by 5.2 % to 5.7 %. Therefore, road charging revenue

redistribution to the private household sector promotes the switch from car to use of the

public transportation. Nevertheless, taking the kilometres travelled in the car or in the

modes of public transport as a homogenous “mobility” bundle, the net effect of road

104

charging on travel activity is in general negative. As a result overall household mobility is

reduced by 4.3 % to 3.9 % in distance travelled depending on the household income

category.

In nominal terms, where the foreign price level is used as numeraire, gross domestic

product (GDP) experiences a positive growth after the implementation of the policy

scenarios. Since with the introduction of the new service “environment” this factor of

production is now explicitly paid for GDP increases. Furthermore, the consumption in

terms of GDP includes only market goods. However, paying a road charge increases the

environmental consumption and this takes place at the expense of traditional consumption.

Therefore the nominal increase in GDP are in line with the decline in domestic

consumption by the private household sector discussed in Chapter 5.3.5.4.

The aggregated welfare calculated within the model quantifies the social benefit from the

reduction of negative externalities from car use. Its level is based on an average external

costs per kilometre calculation as an approximation of marginal external transport costs.

The aggregate external cost is calculated by multiplying the average marginal costs with the

total car kilometres travelled under the assumption of a linear relationship between the

monetarized level of the negative externality and the car road kilometres made.

Assuming

linearity of external costs in distance travelled is a simplification. In reality most functional

relationships between car use and externality generation are non-monotonous, and non-

linear. External effects from fuel consumption, e.g., emissions of CO, HC, NOx, CO

, but

also accidents, etc. are not only a function of the distance driven but also of vehicle speed,

the technical characteristics of the vehicle, driving behaviour, and other factors.

Nevertheless, to some extent the assumption of linearity between the external damage and

car travel is legitimate when costs are aggregated over a large population, partially

balancing out different non-linearities (Small and Kazimi, 1995). The resulting net welfare

benefit is highest for the scenario without the transfer of the road charging revenues to

In the approximated external cost calculation, the following categories are taken into account: infrastructure

depreciation costs, external accident costs, and environmental costs (noise, local pollutants, climate effects),

each differentiated by type of street and user, and net of public revenues raised, e.g., from taxes on insurance,

vehicle registration and fuels (Herry and Sedlacek, 2003; Infras/ IWW, 2000 and 2004). For more details on

marginal social cost pricing approaches see Chapter 2.1.

105

private households since the scenarios with transfer induce lower car use reductions. As

external transport costs in fact can be assumed to rise progressively with transport volume

rather than linearly as is approximated here, the benefit quantification can be considered

rather conservative (Steininger et al., 2007).

5.3.2 Environment

The model implementation allows quantifying the reduction of fuel use related transport

externalities, i.e. CO

emissions, depending on the scheme design. Thus, due to the

reduction in car travel, CO

emissions generated in the motor vehicle sector go down on

average by about 6 %, depending on the revenue redistribution policy. The positive

environmental effect of CO

reduction is based on the fact that the average CO

emission

per passenger-km is far lower for public transport use than for car travel (VDV-Statistik,

2006). Therefore, because of the sizeable reduction in car use and despite of the rise in the

use of public transport due to the modal shift, the overall CO

emission level declines.

Overall results demonstrate that when the ultimate policy objective is a reduction in the fuel

combustion externalities from car use, direct revenue transfer to private households should

be minimized to avoid a response similar to a rebound effect. Results on individual

household contribution to the reduction on CO

emissions are further elaborated in

Chapter 5.3.5.3.

5.3.3 Sectors

Results obtained for selected economic sectors correspond to the results presented in Table

8 for transport related variables and macroeconomic indicators. As one would expect, the

economic activity in sectors related to car travel demand decreases with the introduction of

car road charging. The most significant decline can be observed for the sectors car

manufacturing (i.e., transport equipment), retail activity (i.e., trading), market services such

as repair, and foremost production of refined petroleum products. On the other hand,

106

sectors related to the positively affected public transport demand and the use of road pricing

revenues for qualitative improvement or a quantitative extension of the road infrastructure

supply or public transport services exhibit higher production. These sectors are mainly

construction, non-market services, or the land transport sector. The transfer of the RC

revenues to sectors such as construction, transport or market services implies an investment

and therefore a qualitative improvement or a quantitative extension of the road

infrastructure supply or public transport services. Also sectors linked to the economic

activity of the road charge collector agency, e.g., electrical goods or the banking and

finance sector, increase their output.

5.3.4 Budget

The positive sector effects merge into a positive impact on GDP (see Table 8). The

decrease in production lowers indirect tax revenues and exerts a downward pressure on

employment. Nevertheless the negative impact on employment is outweighed by the

positive labour market effect triggered by the sector shift in production. Therefore, the

economy experiences after all a positive effect on indirect taxes that results from the

positive labour market effect and the moderate effect on private household welfare (see

Table 8 and Table 28). In summary, shrinking public tax revenues are compensated for by

the (semi-public) net revenues collected from car road pricing.

5.3.5 Microeconomic impacts

The introduction of an overall road use charge on car travel at the level of 0.05 Euro per km

means a considerable increase in the unit price of car travel compared to the variable car

km cost before road use charging (on average of about 0.08 Euro) shown in Table 9.

107

Table 9 Car and public transport expenditures across household categories

Transportation cost in Euro per km for different household groups and different transportation means, Germany 2002

Residential category

Public transportation cost in [Euro/km]

Income quartile HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ1 0.09 0.11 0.08 0.10 0.10

HIQ2 0.08 0.09 0.05 0.04 0.07

HIQ3 0.05 0.06 0.05 0.06 0.05

HIQ4 0.07 0.07 0.07 0.05 0.07

Total 0.07 0.08 0.06 0.06 0.07

Car variable cost in [Euro/km]

HIQ1 0.07 0.06 0.08 0.05 0.06

HIQ2 0.07 0.09 0.08 0.09 0.08

HIQ3 0.08 0.08 0.08 0.09 0.08

HIQ4 0.07 0.07 0.08 0.09 0.07

Total 0.07 0.08 0.06 0.06 0.08

Sources: MiD 2002 (infas and DIW Berlin, 2002), EVS 2003 (StaBuA, 2005), and own calculations.

The resulting effects vary considerably across household income groups and with reference

to the road charging revenue redistribution policy scenario as will be presented in Chapters

5.3.5.1 and 5.3.5.4.

5.3.5.1 Household travel expenditure

After the introduction of 0.05 Euro per km of car road charging, expenditure for car travel

as well as for public transportation use increases with respect to the pre-policy situation for

each household group, irrespective of the policy scenario introduced. In Table 10 transport

expenditure impacts from car road charging are presented for different household groups as

%-change relative to the reference scenario for different revenue redistribution schemes.

108

Table 10 Car and public transport expenditure impacts across household

categories and road charging revenue reallocation scenarios

Distributional impacts across road pricing policy scenarios and household groups, Germany 2002

Transport expenditure impacts from car road charging in % change relative to the reference scenario for different revenue

redistribution schemes

Equal household refund in [%-change]

Car

Public Transport

Income

Category HRCent1

HRAgglo2

HRUrb3 HRRul4 Total

HRCent1

HRAgglo2

HRUrb3 HRRul4 Total

HIQ 1 13.5 13.0 14.1 23.3 15.6

4.7 4.5 5.9 9.8 5.3

HIQ 2 14.1 10.9 13.7 12.4 12.5

5.9 4.7 5.9 5.5 5.4

HIQ 3 12.5 13.5 15.5 14.1 13.8

5.3 5.7 6.8 6.4 5.8

HIQ4 14.7 13.8 12.7 12.2 13.6

6.2 5.9 5.6 5.6 5.9

Total 13.8 13.1 13.8 13.9 13.5

5.5 5.2 6.0 7.1 5.6

Proportional household refund in [%-change]

HIQ 1 13.1 12.5 13.5 22.1 14.9

4.3 4.0 5.3 8.7 4.3

HIQ 2 13.9 11.0 13.7 12.1 12.5

5.7 4.8 5.9 5.3 5.7

HIQ 3 12.4 13.8 15.5 13.7 13.8

5.3 6.0 6.9 6.1 5.3

HIQ4 14.8 14.1 12.8 11.8 13.8

6.3 6.2 5.7 5.3 6.3

Total 13.8 13.4 13.8 13.5 13.6

4.0 4.8 6.0 6.2 5.5

No household refund in [%-change]

HIQ 1 12.8 12.2 13.0 21.6 14.5

4.0 3.8 4.9 8.2 4.4

HIQ 2 13.4 10.6 13.1 11.5 12.0

5.2 4.3 5.4 4.7 4.9

HIQ 3 12.0 13.3 15.0 13.1 13.3

4.8 5.5 6.3 5.5 5.4

HIQ4 14.4 13.7 12.2 11.2 13.3

5.9 5.8 5.2 4.7 5.7

Total 13.3 12.9 13.3 12.9 13.1

4.9 4.9 5.4 6.0 5.1

Sources: GRTPM, and own calculations.

Looking at all three policy scenarios, the increase in expenditures for car travel is much

(two to three times) higher than for public transportation use. This mainly results from the

limited substitutability of public transportation journeys for car trips. Moreover, with the

redistribution of road charging revenues to the private household sector the increase in

expenditure for travel related goods and services is higher than without the refund.

Expenditure changes also differ with household category and the differences are most

pronounced for specific combinations of households by income and residential location.

Irrespective of the policy scenario, i.e. with or without introduction of some form of

109

household refund, households in the lowest income quartile experience the highest increase

of car use expenditures and the lowest increase in public transportation spending. The

highest income group experiences the lowest relative increase for car travel. Therefore,

road charging works regressively on car use expenditures.

In contrast, changes in public

transportation expenditures are clearly progressive across household income quartiles, i.e.

they constitute a greater proportion of income as income rises. Thus, households in the

lowest income quartile experience the lowest and households in the highest income quartile

the highest change in public transportation expenditure. When residential location is

considered as single distinction, results turn out rather homogeneous between household

groups and their interpretation is therefore not straightforward.

When household refund is assumed (scenarios A and B) households in the two lowest

income quartiles benefit most from the revenue transfer in terms of increasing spending on

travel activities. Furthermore, while expenditure impacts vary between 12 to 15.6 % for car

use and 4.4 to 6.1 % for public transportation when income and residential location

characteristics are considered separately, they diverge between 10.6 and 23.3 % and 3.8 and

9.8 % respectively when household categories by both income and residence are taken into

account.

For spatially and income disaggregated household categories, households in the lowest

income quartile residing in a rural area experience by far the highest car expenditure

increases, ranging from 21.6 % for scenario C to 23.3 % when scenario A is assumed.

Therefore, for this household group the introduction of road charging has a clearly

regressive impact on their car use expenditures. The same effect is also observed for the

public transportation spending of this household group. On the other hand, households

living in city centres display a rather progressive car expenditure change across household

income quartiles. Again, a similar result is also here true for the expenditures on public

The terms regressive and progressive will be used throughout the work to desctribe the income regressivity

or progressibvity effect with regard to expenditures as well as to welfare impacts occuroing after the

introduction of the policy reform, i.e., car road charging. Hence, a progressive effect in expenditure refers to a

situation, where the percentage change according to household income category is higher for high income

quartile groups than for low income quartile housheholds. Accordingly, a regressive distribution refers to an

outcome where low income quartile households experience a higher relative change than high income quartile

categories (Suits, 1977 and Kiefer 1983).

110

transportation.

Most of these important findings can be interpreted based on a range of different factors,

above all to household specific travel expenditure as well as travel demand profiles in the

pre-policy situation resulting among others from household demographics and

motorisation, as will be discussed in more detail in Chapter 5.3.5.2 together with the

description of mobility impacts from road charging.

In the pre-policy situation households display clearly different expenditure patterns for car

use as well as for the use of public transportation, depending on their income level and their

residential location. Table 11 shows selected household income and expenditure parameters

in absolute numbers. An additional table with absolute household expenditures on fixed car

use related components is presented in the Appendix 6.

111

Table 11 Household income, overall consumption and selected transportation

expenditures for household categories

Household net income total consumption and transportation expenditure in Euro, Germany 2003

Residential category

Net income in [billion Euro]

HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ1 60.6 60.7 41.5 27.0 189.8

HIQ2 58.0 109.9 77.1 48.8 293.8

HIQ3 71.0 170.2 68.2 39.0 348.4

HIQ4 101.5 244.8 79.0 38.3 463.6

Total 291.1 585.6 265.8 153.1 1,295.6

Consumption Total in billion Euro]

HIQ1 51.3 51.9 35.1 22.7 160.9

HIQ2 46.7 89.1 62.1 39.0 236.8

HIQ3 56.4 133.6 49.8 28.1 267.9

HIQ4 72.4 173.3 57.0 27.6 330.3

Total 226.8 447.9 204.0 117.3 996.0

Total fixed car travel expenditure in million Euro]

HIQ1 1,780 1,435 1,802 1,232 6,249

HIQ2 4,002 7,874 6,084 4,094 22,054

HIQ3 5,707 13,525 5,036 3,179 27,446

HIQ4 8,241 20,131 8,451 3,952 40,775

Total 19,730 42,965 21,372 12,456 96,524

Total variable car travel expenditure for fuels in million Euro]

HIQ1 706.2 501.5 817.2 580.4 2,605

HIQ2 1,625 3,073 2,485 1,643 8,826

HIQ3 2,104 5,175 2,492 1,441 11,211

HIQ4 2,916 7,709 2,824 1,471 14,920

Total 7,351 16,458 8,618 5,135 37,562

Total car travel expenditure in million Euro]

HIQ1 2,486 1,937 2,619 1,812 8,854

HIQ2 5,626 10,947 8,569 5,737 30,880

HIQ3 7,811 18,700 7,527 4,619 38,657

HIQ4 11,157 27,839 11,275 5,423 55,694

Total 27,081 59,423 29,990 17,592 134,085

Public transportation in million Euro]

HIQ1 1,152 1,035 472.1 253.9 2,913

HIQ2 608.9 725.4 377.5 211.8 1,924

HIQ3 614.5 920.6 273.9 164.6 1,974

HIQ4 848.5 1,276 270.1 148.8 2,543

Total 3,224 3,957 1,394 779.2 9,353

Total travel expenditure in million Euro]

HIQ1 3,638 2,972 3,091 2,066 11,767

HIQ2 6,235 11,673 8,946 5,949 32,803

HIQ3 8,426 19,620 7,801 4,784 40,631

HIQ4 12,006 29,115 11,545 5,571 58,237

Total 30,304 63,380 31,384 18,371 143,438*

Sources: EVS 2003 (StaBuA, 2005), and own calculations. *Total travel expenditure excludes aviation and holiday travel packages.

112

For the interpretation of the expenditure impacts from road charging it is equally useful to

look at relative expenditure shares in household income as shown in Table 12 and

additionally in households’ overall consumption expenditure as presented in Appendix 7.

Additional information on household expenditures for fixed car use related components as

shares in household income can be found in the Appendix 8. In the Appendix 9 shares of

fixed car use related expenditure components in the overall consumption expenditure are

presented.

113

Table 12 Household overall consumption and selected transportation

expenditures as shares in household net income for household categories

Total consumption and transportation expenditures as net income shares in % for different household groups, Germany 2003

Residential category

Total consumption expenditure in [%-change]

Income

quartile HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ1 84.6 85.5 84.6 83.9 84.8

HIQ2 80.5 81.0 80.5 79.9 80.6

HIQ3 79.5 78.5 73.0 72.1 76.9

HIQ4 71.4 70.8 72.2 72.1 71.2

Total 77.9 76.5 76.7 76.7 76.9

Total fixed car travel expenditure in [%-change]

HIQ1 2.9 2.4 4.3 4.6 3.3

HIQ2 6.9 7.2 7.9 8.4 7.5

HIQ3 8.0 7.9 7.4 8.2 7.9

HIQ4 8.1 8.2 10.7 10.3 8.8

Total 6.8 7.3 8.0 8.1 7.5

Total variable car travel expenditure for fuels in [%-change]

HIQ1 1.2 0.8 2.0 2.1 1.4

HIQ2 2.8 2.8 3.2 3.4 3.0

HIQ3 3.0 3.0 3.7 3.7 3.2

HIQ4 2.9 3.1 3.6 3.8 3.2

Total 2.5 2.8 3.2 3.4 2.9

Total car travel expenditure in [%-change]

HIQ1 4.1 3.2 6.3 6.7 4.7

HIQ2 9.7 10.0 11.1 11.8 10.5

HIQ3 11.0 11.0 11.0 11.8 11.1

HIQ4 11.0 11.4 14.3 14.2 12.0

Total 9.3 10.1 11.1 11.5 10.3

Total public transportation expenditure in [%-change]

HIQ1 1.9 1.7 1.1 0.9 1.5

HIQ2 1.0 0.7 0.5 0.4 0.7

HIQ3 0.9 0.5 0.4 0.4 0.6

HIQ4 0.8 0.5 0.3 0.4 0.6

Total 1.1 0.7 0.5 0.5 0.7

Total travel expenditure in [%-change]

HIQ1 6.0 4.9 7.5 7.2 6.2

HIQ2 10.8 10.6 11.6 12.2 11.2

HIQ3 11.9 11.5 11.4 12.3 11.7

HIQ4 11.8 11.9 14.6 14.5 12.6

Total 10.4 10.8 11.8 12.0 11.1*

Sources: EVS 2003 (StaBuA, 2005), and own calculations.

*Total travel expenditure excludes household expenditures for aviation as well as for holiday travel packages.

Regarding the distribution of absolute income and consumption expenditure volumes for

household income quartiles and residential location characteristic separately and as

114

combined category remarkable differences can be observed. Households in the lowest

income quartile dispose of less than one half of the net income available to the highest

income quartile, or about 15 compared to 36 % of the overall household income in 2003.

The differences between the second and the third quartile are less pronounced with shares

of 23 and 27 % in the overall household income.

By household residential location, highest income proportion (45 %) is allocated to

households in agglomerations. This corresponds to the high household population share of

about 43 % attributed to this residential location category (see Table 6). The next two big

groups according to available overall income are residents of city centres with 23 % and of

urban regions with 21 %. Households in rural regions comprise of about 12 % of the overall

household population and they dispose of almost 12 % of the overall household net income

available within the economy. For combined household income and residential location

categories the row and column distribution of net income volumes looks more complex.

While the distribution of net income volumes within each household income quartile over

the four residential location groups follows the pattern of household population distribution

and is comparable to the overall income distribution for aggregated household residential

categories, the distribution within each location category and over household income

quartiles is more differentiated and does not correspond exactly to the household population

distribution across these categories (see Table 6). The pattern of clearly progressive income

distribution across income quartiles observed when no residential characteristic is

accounted for can only be found within agglomerations. Within households living in city

centres, a share of 20 % in total household net income volume of this residential category is

available to households in the lowest income quartile compared to the highest share of

34 % in household population falling in this group. From the second quartile on, income

shares are distributed progressively and household population shares are distributed almost

equally within city centres’ residents.

The distribution of household net income shares across income quartiles within urban

regions and rural regions are rather similar except for the highest income quartile. Hence,

while in urban regions almost one third of the overall household income is available to the

115

quartile its share comprises only one fourth in rural regions. In both residential

categories households attributed to the top income quartile comprise of the relatively

smallest household population shares of no more than 20 %.

The distributional patterns of household income and household population volumes are

reflected in the average monthly household net income presented in Table 6. Furthermore,

the distributional pattern of income across household categories implies the structure for

households’ total consumption expenditure shown in Table 11 and Table 12, where overall

consumption shares in net income behave regressively across income quartiles. This

includes household overall expenditure for travel, foremost for car ownership and use

related expenditures, but not so for public transportation.

Hence, households’ income shares spend on overall travel goods and services rise from

6.2 % for the lowest to 12.6 % for the highest income quartile. This observation is in line

with findings on significantly positive income elasticities for car ownership and car use.

The patterns are different for car vs. public travel expenditures. While household income

expenditure shares for car use increase with rising income from 4.7 to 12 %, they decline

for the use of public transportation from 1.5 for the bottom to 0.6 % for the upper income

quartile. Progressivity for car use expenditures and regressivity for public transportation

spending across household income quartiles can be observed to a different extent within

each of the four household residential location categories. Furthermore, as Table 12 shows

expenditure income shares for car use increase with declining population density of the

residential location category, from 9.3 for urban centre dwellers to 11.5 % for households

living in rural areas. This is inline with the overall expectation assuming that households

living in big cities with well developed accessibility are less dependent on their automobile,

while for households resident in less populated, remote rural regions the automobile is often

the only mean to meat their travel needs. This can be equally observed regarding the

structure of expenditure income shares for use of public transportation across household

residential location groups, falling from 1.1 % for households in urban centres to 0.5 % for

For further reading on income elasticities for car ownership and car use see also Dargay and Gately, 1999;

Dargay, 2001; Hanly et al., 2002; Johansson-Stenman, 2002; Pucher and Renne, 2003; Giuliano and Dargay,

2006; Kletzan et al., 2006.

116

those living in rural or small town urbanised areas.

The income spent on road use after the introduction of the 0.05 Euro per car km charge has

similar shares (in income) and comparable distribution across household groups as the

expenditure on car use related fuel tax as shown in Table 13. Table 14 additionally shows

the absolute level of the road use charge spent by each household category.

Table 13 Household expenditures on road charging and fuel tax as share in

income for different household categories

Income shares of fuel tax and road charge expenditures for different household categories, Germany 2002

Residential category

Fuel tax expenditure share in household income in [%-change]

Income quartile HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ1 0.7 0.5 1.2 1.4 0.9

HIQ2 1.8 1.8 2.0 2.1 1.9

HIQ3 1.9 1.9 2.3 2.3 2.0

HIQ4 1.8 2.0 2.3 2.4 2.0

Total 1.6 1.8 2.0 2.1 1.8

Road charging expenditure share in household income in [%-change]

HIQ1 0.8 0.6 1.2 2.0 1.0

HIQ2 1.8 1.4 1.9 1.8 1.7

HIQ3 1.7 1.9 2.2 2.0 1.9

HIQ4 2.0 2.0 2.2 2.1 2.1

Total 1.7 1.7 2.0 2.0 1.8

Sources: MiD 2002 (infas and DIW Berlin, 2002), EVS 2003 (StaBuA, 2005), and own calculations.

117

Table 14 Level of road charging payment and household refund according to

scenario B in million Euro for different household categories

Road charging expenditure and refund for different household categories, Germany 2002

Residential category

Road charging payment in [million Euro]

Income quartile HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ1 486 362 483 531 1,863

HIQ2 1,023 1,542 1,496 882 4,943

HIQ3 1,240 3,230 1,471 793 6,734

HIQ4 2,078 4,885 1,760 787 9,511

Total 4,827 10,020 5,210 2,993 23,049

Household refund from road charging revenue in [million Euro]

HIQ1 123 87 142 101 453

HIQ2 282 534 432 286 1,535

HIQ3 366 900 433 250 1,949

HIQ4 507 1,340 491 256 2,594

Total 1,278 2,862 1,498 893 6,530

Sources: MiD 2002 (infas and DIW Berlin, 2002), EVS 2003 (StaBuA, 2005), and own calculations.

Hence, the road charging burden as share in household income is strongly progressive

across household income quartiles varying from 1 to 2.1 % and only slightly progressive

across residential categories with decreasing population density and deteriorating

accessibility, changing from 1.7 for city residents to 2 % for inhabitants of less populated,

rural or small urban areas. The progressivity of the road charging distribution across

household income quartiles sustains within the different household residential categories.

Furthermore, compared to the income shares spent by private households on the overall

travel expenditure varying between 5 and 15 % (see Table 12) an expenditure share on road

use charging between 1 and 2 % can be considered as economically still justifiable.

The examination of absolute road charge amounts paid by household income quartiles as

well as received refund according to policy scenario B

(see Table 14) and corresponding

In scenario 1) 1/3 of the collected road charging revenue is redistributed evenly, i.e. in equal proportions,

across the household categories. In scenario B revenue from road charging is reallocated to the household

categories according to the specific fuel tax burden of each group as proportion of the overall fuel tax burden

contributed by the private households to the state budget (see Appendix 5).

118

shares in the revenue total (see Table 15) show a clearly progressive pattern, where the

lowest income quartile contributes only 8 % to the road charging revenue total compared to

over 40 % paid by households in the highest income quartile. Within the residential

location categories households living in agglomerations – of which most are suburban areas

of large or medium sized cities – pay the highest share (about 44 %) in the overall road

charging revenue. Households in rural areas account for the lowest share of almost 14 % in

the road charge revenue total.

The distribution of absolute amounts of household refund from road charging according to

scenario B as presented in Table 14 follows the relative shares in fuel tax paid by each

household category. Since the road charging payment depends on car kilometres travelled

and therefore on fuel consumption, distributional structures shown in Table 15 are rather

similar. In the case of scenario A each of the 16 household categories receives an equal

refund amount of 408 million Euro or between about 2,000 to 12,000 Euro per household

depending on the number of households in each category as shown in Table 6.

119

Table 15 Household share of the road charging payment and the policy refund in

the revenue and refund total according to scenario B in % for different

household categories

Distribution of road charging payment and revenue for different household categories

Residential category

%-distribution of road charging payment across household categories

Income quartile HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ1 2.1 1.6 2.1 2.3 8.1

HIQ2 4.4 6.7 6.5 3.8 21.4

HIQ3 5.4 14.0 6.4 3.4 29.2

HIQ4 9.0 21.2 7.6 3.4 41.3

Total 20.9 43.5 22.6 13.0 100.0

%-distribution of household refund from road charging revenue across household categories for the policy

scenario B

HIQ1 1.9 1.3 2.2 1.5 6.9

HIQ2 4.3 8.2 6.6 4.4 23.5

HIQ3 5.6 13.8 6.6 3.8 29.8

HIQ4 7.8 20.5 7.5 3.9 39.7

Total 19.6 43.8 22.9 13.7 100.0

Sources: MiD 2002 (infas and DIW Berlin, 2002), EVS 2003 (StaBuA, 2005), and own calculations.

For the comparison of road use charging burden vs. household refund from the road

charging revenue Table 16 presents the individual household refund shares in income for

different household groups depending on the road charging revenue redistribution policy.

120

Table 16 Road charging refund redistributed to private households as share in

income, different household categories

Household road charging revenue refund %-share in income for different policy scenarios and different household

categories

Household road charging revenue refund %-share in road charging payment for different policy scenarios and

different household categories

Residential category

Scenario A in [% of total]

Income quartile HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ1 84.0 112.7 84.5 76.8 87.6

HIQ2 39.9 26.5 27.3 46.3 33.0

HIQ3 32.9 12.6 27.7 51.5 24.2

HIQ4 19.6 8.4 23.2 51.8 17.2

Total 33.8 16.3 31.3 54.5 28.3

Scenario B in [% of total]

HIQ1 25.3 24.0 29.4 19.0 24.3

HIQ2 27.6 34.6 28.9 32.4 31.1

HIQ3 29.5 27.9 29.4 31.5 28.9

HIQ4 24.4 27.4 27.9 32.5 27.3

Total 26.5 28.6 28.8 29.8 28.3

Sources: MiD 2002 (infas and DIW Berlin, 2002), EVS 2003 (StaBuA, 2005), and own calculations.

5.3.5.2 Household travel demand

Selected descriptive characteristics of disaggregated household categories (economic, travel

behaviour related, and sociodemographic) provide substantial insight into different mobility

patterns and the explanation of the corresponding effects from road charging. As shown in

Table 6, whereas by definition the number of private households is equally distributed over

the four income quartiles, the distribution of income over the different household quartiles

as well as the residential location categories is far less evenly distributed as discussed in

Chapter 5.3.5.1 and shown in Table 11. Between income and the number of kilometres

travelled by car in each household category a clearly progressive relationship can be

observed (see Table 18 and Table 19).

123

Table 18 Car, public transportation, and total distance travelled in km per year

and per household across household categories

Household travel activity across household categories, Germany 2002

Total km per year in billion Km per household and per year

Residential category

Car travel

Income

quartile HRCent1 HRAgglo2

HRUrb3 HRRul4

Total HRCent1

HRAgglo2

HRUrb3

HRRul4

Total

HIQ 1 10.7 8 10.5 12 41.1 3,431

2,468

5,083

8,944

4,215

HIQ 2 22 32.5 31.9 18.7 105.2 11,162

9,455

13,339

12,490

11,305

HIQ 3 26.4 68.9 31.5 16.9 143.7 13,784

14,350

17,597

16,417

15,069

HIQ4 44.7 104.2 37.3 16.6 202.8 20,278

20,943

24,095

21,063

21,312

Total 103.8 213.7 111.2 64.1 492.8 11,263

12,983

14,277

13,792

12,930

Public transportation travel

HIQ 1 12.5 9.3 6.3 2.4 30.5 3,998

2,864

3,043

1,817

3,122

HIQ 2 7.9 8.1 8.1 5.2 29.5 4,027

2,363

3,394

3,508

,165

HIQ 3 13.5 15 5.5 2.8 36.8 7,038

3,130

3,052

2,699

3,854

HIQ4 12.3 17.2 3.8 3.2 36.5 5,570

3,458

2,485

4,018

3,835

Total 46.2 49.6 23.7 13.6 133.1 5,012

3,016

3,042

2,930

3,494

Overall household travel

HIQ 1 23.2 17.3 16.7 14.4 71.6 7,428

5,332

8,126

10,761

7,336

HIQ 2 30 40.7 40.1 23.9 134.6 15,189

11,818

16,733

15,999

14,470

HIQ 3 39.9 83.9 37 19.6 180.5 20,822

17,480

20,649

19,116

18,923

HIQ4 56.9 121.4 41.2 19.8 239.3 25,848

24,401

26,580

25,083

25,147

Total 150 263.3 134.9 77.7 625.9 16,275

15,999

17,319

16,722

16,424

Sources: MiD 2002 (infas and DIW Berlin, 2002), and own calculations.

124

Table 19 Distribution of car, public transportation, and total distance travelled

across household categories

Distribution of car, public transportation, and total distance travelled across household categories in %, Germany

2002

Residential category

Car travel in [% of total]

Income

quartile HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ1 2.2 1.6 2.1 2.4 8.3

HIQ2 4.5 6.6 6.5 3.8 21.3

HIQ3 5.4 14.0 6.4 3.4 29.2

HIQ4 9.1 21.1 7.6 3.4 41.1

Total 21.1 43.4 22.6 13.0 100.0

Public transportation travel in [% of total]

HIQ1 9.4 7.0 4.7 1.8 22.9

HIQ2 6.0 6.1 6.1 3.9 22.1

HIQ3 10.1 11.3 4.1 2.1 27.6

HIQ4 9.2 12.9 2.9 2.4 27.4

Total 34.7 37.3 17.8 10.2 100.0

Total travel activity in [% of total]

HIQ1 3.7 2.8 2.7 2.3 11.4

HIQ2 4.8 6.5 6.4 3.8 21.5

HIQ3 6.4 13.4 5.9 3.1 28.8

HIQ4 9.1 19.4 6.6 3.2 38.2

Total 24.0 42.1 21.6 12.4 100.0

Sources: MiD 2002 (infas and DIW Berlin, 2002), and own calculations.

This relationship is valid for household annual total car kilometres as well as for car

kilometres travelled per household and per year. Nevertheless, while the highest income

quartile comprises almost 2.5 times the income of the lowest household quartile,

households in this category make 5 times more car kilometres per year (overall and per

household) than households in the bottom income group.

According to residential location characteristics, households in the rural area display the

lowest annual car use intensities of about 64 million km compared to 214 billion km driven

by households in agglomerations. This distribution can be partly explained by the number

of households in each residential category. When annual car kilometres per household are

125

calculated, households in large city centres have the lowest number in car kilometres

travelled per year (11,263 km). The highest per household car use intensities in km have

households in urban areas, often living in outer suburbs and having to commute to work

longer distances. This distribution of car use intensities is in line with overall expectations,

assuming that people residing in big cities have better accessibility than those in rural area

or small-town urban regions, where commuting to work to larger neighbouring cities is

often the case.

The assumption about poor accessibility found in rural areas is further confirmed by the

distribution of public transportation kilometres travelled by households, as annual total as

well as per household and per year (see Table 18). Hence, highest public transportation use

intensities (as annual household totals in km) are found for households living in

agglomerations and city centres and lowest for those living in rural areas. Public

transportation kilometres made per household and per year underline the regional

differences in mode specific household travel profiles, where households from big cities

travel almost twice as many kilometres per household and per year (5,012 km) by public

transportation means as do households in rural areas (2,930 km).

Regarding the distribution of public transportation kilometres travelled by household

income quartiles, the progressivety between income and travel observed for car use is not

observed, in particular when per household kilometres are taken into account (see Table 19

and Table 25). Furthermore, the differences in overall or per household distance travelled

by public transportation according to income quartile are much more pronounced than for

car travel. Car use intensities are more variable across household income categories than

kilometres travelled by means of public transportation.

Even though it can be assumed that households in the top income quartile are less affine to

take public transportation means compared to those in the bottom income category,

households in the top income quartile are more likely to consist of more members including

children who in general are more prone to use public transportation to reach school or other

educational institutions. Table 6 shows the increasing average household size with rising

household income. Moreover, households with children are in general more mobile than

126

families without children. Table 20 shows overall daily household travel per household for

different household types – with and without children – and different trip purposes.

Table 20 Overall kilometres travelled including all modes per day and per

household by household type and trip purpose

Overall kilometres travelled per day and per household by household type and trip purpose, Germany 2002

Trip purpose in [km/ day/ hh]

Household type Commuting

Education

Business

related Accompany

Private

business Shopping Leisure Total

Working single 14 0 11 1 3 3 18 51

Not-working single 1 1 0 1 4 3 12 21

Single parent 17 7 8 9 7 7 33 88

Couple without

children both working

32 0 26 2 6 8 30 105

Couple without

children one working

15 2 12 2 8 8 30 77

Couple without

children not-working

1 1 0 2 10 9 29 51

Couple with children

both working 32 7 24 13 9 11 53 149

Couple with children

one working 25 6 17 16 9 12 52 136

Other household type

33 10 21 7 11 14 55 151

Average over all

household types 16 3 12 4 7 8 31 81

Sources: MiD 2002 (infas and DIW Berlin, 2002), and own calculations.

Hence, couples with children or single parents on average display higher daily travel

activities. Couples without children make per day from 51 to 105 km depending on their

employment status. However, couples with children travel between 136 and 140 km per

day. The reason is that children themselves have travel demands and at the same time

generate extra household mobility since very often they need to be accompanied by older

individuals or adults, mostly other family or household members. The differences in daily

127

overall distance travelled according to household type and trip purpose as presented in

Table 20 are characteristic for all settlement categories.

Comparing household type residential location distribution according to the presence of

children in the household it can be seen that despite the dominating share of the overall

households in Germany living in agglomerations and urban regions, families with children

or multiple-member households are stronger represented in suburban and peripheral, or

rural regions as shown in Table 21.

Table 21 Household distribution as to residential location and the presence of

children in the household

Household distribution as to residential location and the presence of children in the household in %, Germany 2002

Residential category in [% of total]

Household type HRCent1 HRAgglo2 HRUrb3 HRRul4

Household without children 72.7 68.4 63.6 63.2

Household with one or more children 16.5 19.8 21.2 22.3

Other household type 10.8 11.8 15.2 14.5

Total 100 100 100 100

Sources: MiD 2002 (infas and DIW Berlin, 2002), and own calculations.

Families with one or more children tend therefore to live in less accessible areas like

suburbs or peripheral regions, where they have the opportunity to reside in a house rather

than a flat. The less intense use of public transport due to lower accessibility in less

populated rural or suburban areas is reflected in the household travel intensities according

to travel mode and residential location attribute presented in Table 18 and Table 19.

Showing the correlation between household income and household size based on the

distribution of single household shares according to household category Table 22 underpins

the assumption that multi-member households are more likely to fall into the upper income

128

quartiles, having higher car use preferences than those in the bottom ones. Thus, the bottom

income quartile displays by far the highest single household rate (63 %). According to

residential location category highest single household rates are found among city dwellers

(41 %).

Table 22 Share in % of single households in the total number of households

according to household category

Share of single households in the total number of households according to household category in %

Residential category in [% of total]

Income quartile HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ 1 64 70 57 55 63

HIQ 2 43 34 30 25 34

HIQ 3 29 41 34 27 36

HIQ 4 19 9 1 4 10

Total 41 35 33 32 36

Sources: MiD 2002 (infas and DIW Berlin, 2002), and own calculations.

Summing up, households falling into higher income categories, who tend to be multiple-

member households with children residing in suburban or rural areas, display higher car use

intensities. Hence, the distribution of public transportation as well as car use intensities by

household category can be largely explained by category specific household size and

household composition.

In general, as shown in Table 19 selected German households appear to be rather

automobile than public transportation use oriented, making annually almost 13,000 km per

household by car and only 3,500 km by public transportation means. Also, the average

workday use of public transportation (12 km) is strongly outweighed by distance travelled

using a car (51 km) as shown in Table 23. Remarkable differences can be also found when

examining the distances travelled per household and per workday according to household

income quartile as well as to the residential location. Variations in household category

129

specific daily travel patterns underpin in general the structures presented in Table 18 and

Table 19.

Table 23 Car and public transportation distance travelled in km per household

and per workday for household categories

Car and public transportation distance travelled in km per household and per workday across household categories

Residential category

Public transportation travel in [km/ hh/ workday]

Income quartile HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ1 12.2 10 10 5.3 10

HIQ2 13.5 9.4 9.7 12.3 10.7

HIQ3 22.1 10.7 10.7 11.8 13

HIQ4 17.7 13.9 11.6 17 14.6

Total 15.7 11.2 10.4 10.6 12

Car travel in [km/ hh/ workday]

HIQ1 14.5 8.3 17.1 23.3 14.4

HIQ2 38.6 35.9 51.8 61.8 43.5

HIQ3 53.6 59.5 67.3 72 61.5

HIQ4 83.6 83.4 100.7 90 86.7

Total 44.4 51.2 56.7 55.9 51.2

Sources: MiD 2002 (infas and DIW Berlin, 2002), and own calculations.

The assumption about an existing “dependency” of some household categories more than

others on car use is in line with single- and multi-motorisation rates (together about 81 %)

compared to the relatively low share of non-motorised households (almost 20 %) presented

in Table 24.

130

Table 24 Household shares according to motorization level for different

household categories

Household %-share in the total number of households according to motorization level across household categories

Residential category

No car in [% of total]

Income quartile HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ1 74.8 86.4 60.5 52.3 72

HIQ2 11.7 4.4 0.7 0.7 4.9

HIQ3 2.1 1 0.1 0 0.9

HIQ4 0 0.1 0 0 0

Total 26.9 17.4 16.2 17.5 19.5

1 car in [% of total]

HIQ1 23.4 12.3 36 41.5 25.4

HIQ2 78.8 90 83.2 76.8 84.1

HIQ3 70.9 73.1 61.9 60.7 68.9

HIQ4 47.9 33.6 17.6 17.4 33.1

Total 52.6 54.1 52.4 51.6 53.1

2 and more cars in [% of total]

HIQ1 1.8 1.2 3.5 6.2 2.7

HIQ2 9.5 5.6 16.1 22.5 11

HIQ3 27 25.9 38 39.3 30.2

HIQ4 52.1 66.4 82.4 82.6 66.8

Total 20.5 28.5 31.4 30.9 27.4

Sources: MiD 2002 (infas and DIW Berlin, 2002), and own calculations.

Moreover, household motorisation and non-motorisation shares across income quartiles and

residential location characteristics exhibit the positive correlation between household

income and car ownership as well as a negative relationship between population density of

residential location, i.e. accessibility potential. By far highest non-motorisation rates are

found in the lowest income quartile (on average 72 %) and for the city centres (about

27 %). In fact, considerable regional differences exist in household mode choice

preferences. In some urban agglomerations more then half of all households do not own a

car, in some rural areas a high share up to 50 % and more of households never has used any

kind of public transportation.

131

The variations in public and car transport demand over the disaggregated household

categories are relevant for the assessment of the acceptability towards road charging

measures since income, the availability of alternatives and household demographics

determine the reaction potential to price increases of car travel (Dargay and Gately, 1999;

Dargay, 2001; Hanly et al., 2002; Giuliano and Dargay, 2006).

Travel mode use differences across household income groups correspond on the whole with

the differences in mode specific expenditure shares by household income, as shown in

Chapter 5.3.5.1. While household expenditure shares on fixed car cost components are

clearly progressive with increasing income, relative expenditures on variable car travel,

mainly fuel, is lowest for the highest income group. This implies that the highest household

income category is relatively less “sensitive” to the introduction of road charges on car use

if they are implemented proportional to distance travelled.

As shown before, expenditure shares for public transport in monthly household income

decline considerably with rising income. Regarding the relative household expenditure

patterns for travel services, the lowest income group has the highest income expenditure

share on public transport, and the lowest on car use. This may by a tentative indicator that

households in this category satisfy their mobility demand as far as possible by public

transport and the mobility demand allocated to private car use cannot be easily shifted to

other transport modes. Real-life examples of such mobility patterns found in Germany are

households living in remote areas of mainly former Eastern Germany with high

unemployment and low population densities, and therefore poor accessibility, where the

only way to get to work is to commute long distances by car. Due to low disposable

income, such households find it often difficult to leave their residential area and move

closer to their economic activity. These households stand as an example for the population

group at risk to experience the highest welfare losses from the introduction of distance

dependent road charging.

Altogether, the structure of mobility parameters presented in this chapter is on the whole in

line with mode specific household travel expenditures presented in Chapter 5.3.5.1. The

initial differences in household expenditure shares on mobility services and the

132

corresponding travel activity parameters reveal the asymmetric availability of reaction

potentials to the pricing measure, and are therefore relevant for the interpretation of the

mobility, but in particular the welfare and equity impacts from car road charges.

Hence, the implementation of road pricing changes the price of car travel and generates a

shift in the modal split resulting in changing overall transport volumes, the extent of which

depends on the reaction parameters introduced in the model. Table 25 presents the effects

the introduction of road use charging has on travel behaviour of different household

categories and according to the policy scenario set in place.

Table 25 Car and public transportation distance travelled impacts across

household categories and road charging revenue reallocation scenarios

Travel behavior impacts across road pricing policy scenarios and household income categories

Impacts on distance travelled from car road charging in % change relative to the reference scenario for different

revenue redistribution schemes

Equal household refund in [% change]

Car km Public transportation km

Income

Category HRCent1

HRAgglo2

HRUrb3 HRRul4 Total

HRCent1

HRAgglo2

HRUrb3 HRRul4 Total

HIQ 1 -8.9 -8.8 -7.0 -10.3 -8.8 4.8 4.5 5.9 9.8 5.3

HIQ 2 -6.9 -5.3 -6.4 -5.4 -6.0 5.9 4.7 5.9 5.6 5.5

HIQ 3 -6.0 -6.5 -6.7 -5.6 -6.4 5.4 5.8 6.9 6.5 5.8

HIQ4 -7.0 -6.6 -5.7 -4.8 -6.4 6.3 5.9 5.6 5.7 6.0

Total -6.9 -6.4 -6.3 -6.2 -6.5 5.6 5.4 6.1 6.5 5.7

Proportional household refund in [% change]

HIQ 1 -9.3 -9.3 -7.5 -11.2 -9.4 4.4 4.0 5.3 8.7 4.8

HIQ 2 -7.1 -5.2 -6.3 -5.6 -6.0 5.7 4.8 6.0 5.3 5.5

HIQ 3 -6.1 -6.3 -6.7 -5.9 -6.3 5.3 6.0 6.9 6.1 5.9

HIQ4 -6.9 -6.2 -5.6 -5.1 -6.2 6.4 6.2 5.7 5.3 6.1

Total -7.0 -6.2 -6.3 -6.6 -6.4 5.4 5.5 6.0 6.1 5.6

No household refund in [% change]

HIQ 1 -9.5 -9.4 -7.9 -11.6 -9.7 4.1 3.8 4.9 8.3 4.5

HIQ 2 -7.5 -5.6 -6.8 -6.0 -6.5 5.3 4.4 5.4 4.8 5.0

HIQ 3 -6.5 -6.7 -7.1 -6.4 -6.7 4.9 5.6 6.4 5.6 5.4

HIQ4 -7.3 -6.6 -6.0 -5.6 -6.6 6.0 5.8 5.2 4.8 5.7

Total -7.3 -6.6 -6.7 -7.1 -6.8 5.0 5.1 5.5 5.6 -5.2

Sources: GRTPM, and own calculations.

133

Irrespective of the revenue reallocation or household refund scheme introduced along with

road charging, each household category reduces its car use and in turn increases its

utilization of public transportation (in km). Comparing these general results with the

impacts on travel expenditures shown in Table 10 it can be observed that despite falling

demand in car kilometres, expenditure for car travel still goes up, basically as a

consequence of low elasticity. In contrast, expenditure for public transportation use rises,

but so does the kilometres demand for the corresponding transportation means.

Considerable differences across household categories exist in travel behaviour reaction to

road charging. Regarding car travel, the highest kilometre reductions are displayed by the

lowest household income quartile (comparing all three policy scenarios between 7 and

11.6 %). Noteworthy, within the lowest income quartile by far the highest reductions in car

use intensities take place in the category of households living in rural areas – depending on

policy scenario between 10.3 and 11.6 %. Income quartiles two to four display rather

moderate car use reductions compared to the other household categories. This can be

explained by the fact that the relative share of motorized travel in overall travel of

households living in remote rural areas with low population densities is already rather high

to begin with, i.e. before the introduction of road charging (benchmark scenario). Poor

accessibility of these regions leads to the assumption that longer car travel distances result

mainly from necessary service trips of different kinds as well as travel to work. Such trips

are often difficult to resign and due to the lack of public transportation alternatives, and

sometimes also due to the length of the trip, they become impossible to substitute by other

modes. Therefore, households in rural areas will react in general with lower car use

reductions to rising car use costs. On the other hand, a radical cut in leisure or other less

necessary trips will take place in households with small budgets in order to afford the

maintenance of the indispensable car trips, such as travel to work, getting medical care, etc.

This can be observed for the bottom income quartile residing in rural regions compared to

the remaining households income groups that show relatively weaker reactions.

Regarding solely the income quartiles, the biggest drop in car use as reaction to road

134

charging takes place when scenario C is introduced (9.7 %). The progressivity at which car

use falls across income quartiles holds for all residential location categories.

The distribution of the negative impact on car kilometres after the introduction of road

charging across households’ residential location categories is more homogenous than across

income quartiles. Irrespective of the revenue redistribution policy car use falls highest in

the category of city residents.

After the introduction of road charging, household demand for public transportation travel

augments between 3.8 and 9.8 %. Interestingly, households in the lowest income quartile

with residential location in a rural area display the highest increases for public

transportation travel, irrespective of the household refund policy. This reaction reflects the

strong budgetary restrictions faced by these households that require the abatement of not

obligatory car trips and the substitution of as many car trips as possible by the use of public

transportation means. On the other hand, households in the bottom income quartile who

live in city centres show on the whole the lowest growth rates in public transportation use

induced by the introduction of road charging. Obviously, households in this category are

only little dependent on car use due to good public transportation coverage generally

available in highly concentrated urban centres. Since most of their trips are already carried

out by modes of public transportation, only limited potential is left for further substitution.

Moreover, because households in this category bear a rather small (variable) car use

expenditure burden due to their overall low use intensities, the impact from increasing car

use costs as consequence from road charging will hit them to a comparatively limited

extent.

As shown in the policy simulation, the redistribution of only a small part of the road

charging revenue to the private households will only moderately mitigate the negative road

charging impact on car use. Therefore, the redistribution of a proportion of the road

charging revenue will not noticeably counteract the environmental objective of CO

(and

NOx) emission reduction. However, a differentiated household refund structure can

significantly absorb the negative net effect on mobility (and welfare) induced by rising cost

of car use (see “Overall travel” in Table 8).

135

5.3.5.3 Households’ contribution to the reduction in CO

emissions

Based on household vehicle ownership information such as engine size, vehicle model and

year of its first registration and the information on corresponding kilometres travelled

available from the MiD survey, household category specific CO

emission parameters were

calculated. The parameters were incorporated in the GRTP model to account for the

observation that households who are better off in terms of income tend to own bigger and

therefore (for the most part) less fuel efficient cars than do poorer households. On the other

hand richer household tend to have newer and therefore more fuel efficient cars. The two

arguments are somehow counterbalancing. However, the assumption of more CO

emission

intensive car travel among richer households cannot be confirmed when household income

quartiles are considered. Hence, while differences in the distribution of CO

emissions in

1,000 t across household categories are quite remarkable because they are mainly driven by

household specific car use, differences in household CO

emissions per car-km as well as

per overall km travelled where household use of public transport is taken into account are

rather small (see Table 26). In line with the positive correlation between household income

and car use, household share related CO

emissions vary with rising income level from

8.7 % for the lowest to 40.8 % for the highest income quartile.

136

Table 26 Selected CO

emission characteristics of household categories in the

pre-policy situation, Germany 2003

emissions from household travel across household categories, Germany 2002

Residential category

Total CO

emission distribution across household categories in [% in total]

Income quartile HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ1 2.4 1.8 2.2 2.3 8.7

HIQ2 4.5 6.7 6.5 3.8 21.4

HIQ3 5.5 14.0 6.3 3.3 29.1

HIQ4 9.2 20.8 7.3 3.4 40.8

Total 21.6 43.2 22.3 12.9 100.0

Total CO

emissions in [kg/overall km travelled]*

HIQ1 0.115 0.114 0.144 0.180 0.134

HIQ2 0.166 0.182 0.179 0.175 0.176

HIQ3 0.152 0.184 0.189 0.188 0.179

HIQ4 0.179 0.190 0.197 0.192 0.189

Total 0.159 0.182 0.183 0.184 0.177

Car CO

emissions in [kg/car km travelled]

HIQ1 0.213 0.211 0.213 0.210 0.212

HIQ2 0.215 0.220 0.217 0.216 0.217

HIQ3 0.215 0.218 0.216 0.214 0.217

HIQ4 0.220 0.216 0.214 0.223 0.217

Total 0.217 0.217 0.215 0.216 0.217

in [1,000 t]

HIQ1 2,659 1,964 2,414 2,583 9,619

HIQ2 4,973 7,403 7,175 4,193 23,744

HIQ3 6,081 15,474 6,974 3,691 32,220

HIQ4 10,193 23,026 8,099 3,797 45,115

Total 23,906 47,867 24,661 14,264 110,698

Sources: MiD 2002 (infas and DIW Berlin, 2002), and own calculations.

*The total CO

emissions are calculated based on (car) driver km for private car transport and on passenger km for public

transport, under the assumption of an average occupancy of 20 person per bus km.

After the introduction of charges for passenger car road use CO

emissions fall mainly due

to decreases in car travel and therefore fuel consumption shown in Table 25. Following the

differences across households in the reduction of car use after the implementation of road

charging, different household categories contribute differently to the decrease of car related

137

emissions as shown in Table 27. According to household equivalence-weighted

income quartile the lowest household income quartile has the highest CO

emission

reduction contributions in each revenue reallocation scenario.

Independently of household income quartile the policy scenario where no private household

refund takes place encompasses the highest CO

declines. In the scenario with equal

household refund shares income quartiles 1 and 4 display highest CO

reductions, reflecting

shifts in car and public transport travel after the introduction of a car road use charge.

Table 27 Overall CO

emission impacts across household income groups and

road charging revenue reallocation schemes

Overall CO

emission impacts from car road charging in % change relative to the reference scenario for different

revenue redistribution schemes, Germany 2002

Residential location

Equal household refund in [%-change]

Income quartile HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ 1 -7.0 -7.0 -6.0 -9.8 -7.5

HIQ 2 -6.3 -5.0 -6.0 -5.0 -5.7

HIQ 3 5.3 -6.1 -6.4 -5.3 -5.6

HIQ4 -6.5 -6.3 -5.5 -4.6 -6.1

Total -6.2 -6.1 -6.0 -5.8 -6.0

Proportional household refund in [%-change]

HIQ 1 -7.4 -7,4 -6.5 -10.6 -8.0

HIQ 2 -6.5 -4.9 -5.9 -5.3 -5.6

HIQ 3 -5.3 -5.9 -6.4 -5.7 -5.6

HIQ4 -6.5 -6.0 -5.5 -4.9 -5.8

Total -6.3 -5.8 -6.0 -6.2 -6.4

No household refund in [%-change]

HIQ 1 -7.6 -7.6 -6.9 -11.0 -8.3

HIQ 2 -6.9 -5.3 -6.4 -5.6 -6.0

HIQ 3 -5.7 -6.3 -6.8 -6.1 -6.0

HIQ4 -6.8 -6.3 -5.9 -5.3 -6.3

Total -6.6 -6.2 -6.4 -6.7

Sources: GRTPM, and own calculations.

138

The distribution of the CO

emission reduction looks less differentiated when household

groups according to residential location are examined.

Overall, results in CO

reductions as effects from the introduction of road charging

demonstrate that when the ultimate policy objective is the reduction in the fuel combustion

externalities from car use, direct revenue transfer to private households should be

minimized.

5.3.5.4 Household distributional welfare and equity effects

The set-up of the GRTPM allows the assessment of welfare that accounts for the changes in

transportation as an economic variable. The welfare measure is based on an agent-based

utility function and therefore all changes in transportation and economic conditions

affecting individuals are incorporated in the welfare measure. In Table 26 welfare impacts

calculated as Hicksian welfare index consisting of all household expenditures and Hicksian

transportation welfare index are reported. The Hicksian equivalent variation indicates the

amount of income necessary to compensate an individual (in the pre-policy situation) in

order to reach equality with the post-policy utility level (Just et al., 2004). It therefore

measures changes in money metric utility between the pre- and post-policy equilibrium,

which is the amount of money required to bring a household back to the same level of

utility as in the benchmark equilibrium following changes in prices in counterfactual

equilibrium. Furthermore, the calculated welfare index does not account for the

environmental welfare improvement but measures only the change in traditional marketed

goods consumption.

As shown in Table 28 the welfare change of unique household categories turns throughout

negative, but varies considerably with the road charging revenue redistribution scenario and

with household category.

Depending on the road charging revenue redistribution policy households in the lowest income quartile

would have to be compensated by a fraction between 0.4 and 1.2 % of their income in order to maintain their

pre-policy overall consumption utility level after the implementation of the road charging policy.

139

Table 28 Welfare impacts across household categories and road charging

revenue reallocation scenarios

Consumption welfare impacts from car road charging in %-change relative to the reference scenario for different

revenue redistribution schemes, Germany 2002

Hicksian welfare index overall household expenditure

Residential location

Equal household refund in [%-change]

Income quartile HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ 1 -0.3 -0.1 -0.4 -0.8 -0.4

HIQ 2 -1.0 -0.9 -1.3 -0.9 -1.0

HIQ 3 -1.0 -1.4 -1.2 -0.7 -1.2

HIQ4 -1.2 -1.4 -1.3 -0.7 -1.3

Total -1.0 -1.2 -1.1 -0.8

Proportional household refund in [%-change]

HIQ 1 -0.8 -0.6 -1.0 -1.7 -0.9

HIQ 2 -1.2 -0.8 -1.3 -1.1 -1.1

HIQ 3 -1.1 -1.1 -1.2 -1.1 -1.1

HIQ4 -1.2 -1.1 -1.2 -1.0 -1.1

Total -1.1 -1.0 -1.2 -1.2

No household refund in [%-change]

HIQ 1 -1.0 -0.8 -1.4 -2.2 -1.2

HIQ 2 -1.6 -1.3 -1.8 -1.6 -1.5

HIQ 3 -1.5 -1.5 -1.7 -1.5 -1.6

HIQ4 -1.5 -1.5 -1.6 -1.5 -1.5

Total -1.4 -1.4 -1.6 -1.7

Sources: GRTPM, and own calculations.

In general, negative overall household consumption welfare changes are rather small,

depending on the policy scenario ranging between 0.3 for households in the bottom income

quartile living in big cities when scenario A is implemented and 2.2 % for those in the

bottom income quartile and in a rural area in case of scenario C. Examining overall

consumption welfare changes according to income quartiles after the introduction of road

charging none of the policy scenarios encompasses a clearly regressive distributional

impact, where the bottom income quartile bears the highest negative welfare burden from

the implementation of the measure. For scenario A the opposite is clearly the case, where

imposing a road charge induces a strongly progressive effect, reducing welfare from 0.4 %

for the bottom to 1.3 % for the top household income quartile. Even though, negative

welfare changes across income quartiles in scenario B and C are not straight forward to

characterise as to their distributional impacts, especially the results obtained for scenario B

140

display an interesting pattern of close to equal welfare losses across the household income

categories. Moreover, for scenario B the only regressive welfare distributional outcome can

be observed across household income quartiles within the rural residential category.

Remarkable differences in welfare losses can be observed between scenarios A and B

compared to scenario C where no household refund from road charging revenue takes

place, in particular for the lowest income quartile(s), but also for households in rural areas.

Each household category is better off when household transfers are introduced.

Nevertheless, the gains from the revenue redistribution towards the private household

sector are most pronounced in the bottom income quartile. With regard to possible equity

implications drawn from the welfare impacts presented in Table 28, scenario B yields the

most evenly distributed welfare losses within the household income or residential location

categories.

The interpretation of overall welfare losses across the four residential location categories as

well as across households disaggregated by income and residential location are not

straightforward. For residential location categories the differences between the groups and

within each policy scenario are less pronounced than for income quartiles. Nevertheless,

when no direct refund is introduced households experience the highest welfare losses when

solely their residential location is taken into account. The introduction of household

transfers reduces the welfare losses according to residential location considerably, foremost

for households living in rural (1.7 compared to 0.8 %) as or small town, urban areas (1.6

compared to 1.1 %) of the country.

Regarding household categories disaggregated with respect to both substantial differences

exist between the row- and column averages and we observed mixed results. The

progressivity of the distribution of the negative welfare effects across income quartiles

observed for scenario A cannot be easily transferred onto the income quartiles within each

residential location group. In the residential location categories 1 through 3 a “close to” or

“interrupted” progressivity across income quartiles can be observed (see Table 28).

Furthermore, welfare losses across income quartiles within the 4

residential location

category for scenario A are very similar, compared to the same categories for scenarios B

141

and C.

All observation discussed are important for the design of possible refund policies with the

introduction of car road charging. Results obtained based on scenario C show which

household categories are to what extent at risk to bear the highest negative welfare burden

from the implementation of road charging. In Germany households at risk have a low

income (bottom quartile) and live in rural areas or have a moderate income (2

quartile)

and live in small town, suburb areas. Nevertheless, despite the considerable welfare losses

among better situated households (top income quartiles) one can argue if these households

are most affected by road pricing. The interpretation of corresponding results is not

straightforward. In general they are in line with the findings that car use increases with

rising income, even though it is equally important to take into account the residential

location of the household to draw conclusions about the distributional and equity effect of

car road charging in Germany. Findings obtained for Germany from the implementation of

the GRPM are in line with results from studies conducted for European countries arguing

that those with low incomes would gain the most from different kinds of road use charging

(see acceptability discussion in Chapter 2.2). This is mainly explained through the fact that

in the European travel context car is often not the dominant mode of transport, in particular

for citizens living in bigger cities who have access to public transportation and are provided

with good conditions for using the so called ‘‘slow modes’’, i.e., cycling and walking.

Assuming that those using the fast mode (car) usually are the more affluent travellers, car

road charging will be progressive (Glazer and Niskanen, 2000; Evans, 1992). Since low-

income groups more often use public transport, not only will they be less affected by the

charges, but they will also profit more from the revenues if they are spent on improving

public transport as designed in the implemented policy scenarios for part of the revenues.

Furthermore, the distributive effects illustrated in Table 28 reproduce on the whole the pre-

policy mobility profiles of the different household income categories and therefore their

“vulnerability” to the road charging policy. The results show that the negative welfare

effect from road charging can be best compensated by revenue redistribution directly to

households. The relatively highest benefit from revenue redistribution is allocated to the

142

bottom income category. An important implication of this result is that the net welfare or

equity effect, and therefore the social acceptance of road charging policies is clearly linked

to the redistribution scheme of the revenues from the measure. This is in line with research

done by Small (1992) showing that different use of revenues will imply different net

effects, and accordingly will determine whether the road use charging policy as a whole

will be progressive or regressive.

As shown in the simulation study, the redistribution of only a small part of the revenue to

specific household groups will induce a rather moderate positive effect on their (car) travel

demand and therefore not counteract the environmental objectives of the measure.

However, a differentiated household refund structure can significantly absorb the negative

welfare effect of the rising cost of car use.

5.3.5.5 Equity measurement

In the economic theory diferent approaches exist to assess income or welfare distributional

as well as equity or inequality issues. Most often some kind of income or welfare function

is defined to apply a summary statistics for comparing different frequency distributions in

order to conclude with a result on inequality measurement. However, the choice of the

inequality or distributional measurement approached depends on the initial data (-

distribution) used to conduct the analysis. Some examples for common empirical

approaches to income or welfare distribution or inequality assessment are the use of ex ante

or ex post (microeconometric) microsimulation modelling using a variety of indices

depending on the underlying data and the specific research question of interest. In general,

within the literature on the measurement of income distributional effects from policy

measures – very often taxation reforms – four different subconcepts can be outlined: first

being the traditional concept of inequality, second being the rather novel concept of

polarisation, third representing the concept of progression in taxation, and fourth

concerning the concept of income poverty. The basis for the development and application

of distributional and inequality indicators is the introduction of an appropriate concept of

income. The appropriate definition of the underlying income is crucial for the interpretation

143

of the results from the distributional analysis. The most prominent descriptive indices of

inequality – mainly based on econometric calculations using descriptive statistics – are,

e.g., the Gini coefficient, the relative mean deviation, the coefficient of variation, the

logarithmic variance, the variance of the logarithms, the Mehran index, and the Piesch

index.

Another group of indices of inequality is derived based on the concept of probability of the

occurrence of events that is based on information theory.

A special and rather normative group of inequality indices is concerned with the concept of

social welfare, where the welfare analysis of distributional effects takes into account

individual preferences, coherent utility functions, the formulation of riskiness, and the

concept of risk aversion. The underlying social welfare function provides the link between

welfare theory and inequality measurement, as they become a function of the equity of an

income distribution. Using the concept of inequality aversion it is assumed that social

welfare increases the more equal incomes are distributed. Examples of welfare

measurement indices are e.g., different formulations of the Atkinson welfare index.

Given the complexity of data base definition and theoretical justification of the

distributional measure to be applied, the calculation of a critical number of even a small

number of distributional indicators goes beyond the data available in this work. In this

study only household income quartiles were constructed, based on the equivalence-

weighted household incomes. This provides a critical number of only four observations for

a possible application of inequality or distributional measures. The reason quartiles rather

than quintiles or deciles were chosen for the model database was to limit the number of

household categories after the inclusion of the four residential characteristics. The

construction of two-dimensional household categories based on, e.g., income deciles and

the four residential location attributes would lead to 40 different household categories with

a likely critical number of observations in each category. Besides, the interpretation of

distributional effects between 40 two-dimensional categories could easily become fuzzy

and unsound. On the other hand, the calculation of distributional indices based on the four

A detailed description and the mathematical derivation of the indices can be found in Ochmann and Peichl

(2006).

144

measurement points used in this work is likely to be critical as far as the interpretation,

comparison, and the reliability of the results is concerned. However, this aspect remains a

crucial point concerning the definition of future research resulting from the results obtained

from this thesis (see Chapter 6).

145

6 Conclusion

The objective of this dissertation work is the (integrated) quantitative impact assessment of

car road charging on welfare and travel demand across private household categories as well

as on selected economic and environmental indicators.

To be able to investigate the research question a computable general equilibrium model for

Germany was constructed. The CGE model for Germany (GRTPM) and the database were

built on the basis of an existing standard model code. The model was furthermore extended

through the inclusion of different household categories. Differentiated household categories

as to equivalence-weighted income and residential location were introduced to allow the

assessment of distributional and equity effects from the implementation of road charging

policies within the model based economic framework. Household categories were specified

through individual road and public transportation demand profiles as to distances travelled

and transportation expenditures. Hence, heterogeneous reaction potentials in response to the

policy scenarios within the private household sector were taken into account. Expenditures

on car travel were further disaggregated into fixed cost of car purchase or ownership and

variable costs of car use. After calibration of the model on the new database, effects of

distance dependent road charging policy reforms implemented in the private passenger car

travel sector were calculated. The policy measure introduced is a distance dependent and

time invariant road charge of 0.05 Euro per car kilometre. The analyses of different road

charging revenue recycling schemes served as the basis for the evaluation of welfare,

distributional, and equity impacts from charging road use of private cars.

Depending on the revenue recycling policy accompanying the road charging scheme a

general reduction in car use together with an increase in the use of public transportation

across household categories can be observed. Household consumption and therefore

household welfare decrease independent of the initial pre-policy income level due to the

road charge. Nevertheless, it needs to be emphasised that this effect depends on how far the

internalisation benefits are taken into account and what kind of revenue use policy is

introduced. However, the magnitude of the distributional impact depends on the initial

household travel behaviour and its socioeconomic profile. The top household income

146

quartile experiences the highest welfare losses, irrespective of the road charging revenue

reallocation scheme. The bottom household income quartile reacts more sensitively to the

introduction of a road charging revenue redistribution policy. Hence, when some proportion

of revenue from road charging is transferred towards the private household sector,

households in the bottom income groups experience the relatively highest mitigation effect

of the negative welfare impact from road charging. This implies that negative welfare or

equity effects – especially burdensome for financially disadvantaged households – and

therefore the social acceptance of road charging policies can be determined by the

redistribution scheme of the road charging revenues. This conclusion holds also for

household net welfare effects according to residential location. When no revenue

redistribution takes place, households from rural areas are clearly the losers of the road

charging policy reform. They can be put significantly better off by the implementation of

household refunds financed from the road charging revenues.

The redistribution of the road charge revenue share directly to the private households

lowers the negative effect on car use as well as on household welfare. At the same time it

prevents the switch from car to public transportation. However, the net effect on car travel

of levying a road charge and the redistribution of a part of the revenue is negative. As a

result, overall household mobility is reduced depending on the household income category.

This is an important implication when pursuing environmental objectives. The introduction

of road charging can clearly lead to the reduction of car use whereas the increase in e.g.

fuel taxes might lead to the use of more fuel efficient cars in the first place without actually

reducing car road use. Moreover, the relevance of car fuel efficiency in the case of a

gasoline price increase is ambiguous in its distributional implications, since richer

households tend to have bigger (and thus less efficient) cars, they are also more likely to

have newer and thus more efficient models. The decision to investigate a per-km road

charge, rather than a gas tax or a vehicle-specific charge is also related to the given

probability of implementation of such an instrument in Germany, as well as to the fact that

energy tax on gasoline is already one of the highest compared to other countries in the

European Union, and in particular compared to Germany’s border countries. A further

147

increase of the tax would worsen the already existing adverse effect of “grey fuel imports

and refueling tourism. At the same time it takes into account two important policy changes

already or soon taking place in Germany: Firstly, the recent developments within the

German vehicle taxation scheme with regard to the consideration of CO

emissions; and

secondly, the European Commission regulation to come in 2012 forcing European car

makers to curb down the CO

emissions of the newly vehicle registrations to 120 CO

g/km. Both policy regulations will improve fuel efficiencies of passenger cars, inducing

decreasing fuel demand. It can be therefore expected that in the long run both policy

changes will relax the impact of fuel taxation as policy instrument to regulate car use and

generating state revenue and making it therefore necessary to set up an alternative policy

instrument such as road use charging.

Model results show furthermore that due to the reduction in car travel, the carbon emissions

generated in the motor vehicle sector are also diminished.

The labour market and public demand experience a positive impact after the

implementation of the policy scenarios, whereas the gross domestic product experiences a

negative impact after its deflation with the decreasing purchasing power parities relative to

abroad resulting from the initial increase in the national production. Hence, the nominal

GDP increases by about 1.3 %, the number of unemployed is reduced by about 2

percentage points, leading to a change in the unemployment rate from 9.3 % to 8.9 % in the

reference year 2002.

The research approach used in this study, including the methodology and the database are

unique for Germany. The construction of the database underlying this study as to the

combination of two different micro data sources, including income estimation for the

delimitation of equivalence-weighted household income quartiles can serve as a good

example for future studies of this kind. Furthermore, households included in the data base

were categorised as to their residential location attribute, and therefore including

information on immanent accessibility potentials. As shown through the results obtained

from this work this is an important extension for the evaluation of travel demand related

policy instruments. Thus, the results obtained together with the methodological tool

148

constructed for this study are a new contribution to the literature of CGE modelling with

integrated passenger travel demand for different household categories and the assessment

of distributional and other economic and environmental effects after the implementation of

a policy measure, e.g., road use charging.

Hence, the study results confirm the assumption that negative individual consumption

welfare effects from implementing road user charging can be best compensated by revenue

redistribution towards the private household sector. The redistribution of only a small part

of the revenue will moderately mitigate the negative effect on household (car) travel

demand induced by road charging and therefore not offset the environmental objective of

the measure. However, an adequate household refund structure can significantly mitigate

the negative welfare effect from rising cost for car use, especially for the household groups

at risk, i.e. bottom income quartile and rural areas’ residents. The pre-policy variations in

public and private passenger transport demand across household income categories are

relevant for the assessment of the acceptability towards road charging measures since

income and the availability of alternatives determine the reaction potential to price

increases of car travel. Pre-policy settlement structures, car availability and public transport

use across household income and residential location groups – all interconnected

parameters – have a decisive impact on the distributional impact of a car road pricing

scheme. The quantitative results of welfare or equity shifts induced by road charging imply

to what extent social acceptance towards the policy measure can be expected. In that sense,

the distributional impact assessment suggests how to design an adequate road charging

revenue redistribution scheme in order to improve its social acceptance across different

population groups, since a differentiated household refund structure can significantly

absorb the negative welfare effect from the rising cost for car use.

To have a better understanding of equity and fairness implications derived from

quantification of distributional and welfare effects from road charging in Germany, the

relevant literature concerning current equity and fairness concepts was examined. Varying

reflections on equity and different methods of its assessment were surveyed to subsequently

investigate the possible outcomes of such public programme regarding the changes and

149

effects on equity. In fact, surveyed literature reveals strongly contrasting views on the

question, indicating the difficulty in finding a clear-cut conclusion about the distributional

and equity measure of road charging. Nevertheless, the broad consensus says that case-

adequate or sufficiently progressive revenue recycling schemes can ensure that all income

groups benefit from the measure, even though each road charging policy will still require

its specific evaluation.

Another conclusion is that equity outcomes will in general strongly depend on the design of

the policy instrument as well as on varying travel patterns from socioeconomic differences

of population groups. This argument could be also confirmed from results obtained within

this work. Evaluation of road charging policies needs to consider distributional effects

before and after the implementation of different revenue redistribution schemes, comparing

net welfare surplus with the total distributional effects. Whether the measurable shifts in

welfare distribution are equitable is much more difficult to answer. Even though different

methods for quantifying equity effects exist, the concept of equity, however, is much more

complicated to “grasp”. A multitude of factors influence the perception of fairness and they

vary among different individuals as well as socioeconomic groups. Various concepts of

what could be meant by the terms “equity”, “fairness” or “justice” exist, nevertheless a

universal consensus has not been found. Attempts to determine overall satisfying concepts

are based on qualitative reflections rather than quantitative methods.

Another aspect to be kept in mind concerning the evaluation of the equity question is the

fact that although it might be feasible to design a revenue disbursement package that will

compensate aggregate losses across income classes, it is rather impossible to try to

compensate all losses to all individuals. Therefore, irrespective of the efforts behind the

design of the policy scenario and the revenue redistribution scheme, it is likely that some

travellers will be made worse off as a result of road charges. This is one core aspect

concerning the implementation of road charging that is in the focus of fairness concerns

when discussing such policy reforms. It has been also suggested by scholars that because of

the large number of factors determining the impacts of car road charging, revenue

redistribution cannot solve all equity and fairness concerns. It is unrealistic to believe that a

150

road charging policy program can be designed in a way that is both completely equitable

and administratively feasible.

Nevertheless, researchers can provide policy makers with useful information regarding

equity issues comparing the distribution of costs and benefits for various road use charging

policies and estimating the total amount of revenue available from the measure

implementation. This partially reflects the objective of this work. The information provided

from the implementation of integrated economic impact assessment tools is essential for

determining redistributions that are equitable enough to satisfy relevant social groups (of

concern) and gain their acceptance for the policy reform.

Road charging policies are most often discussed within the context of such concepts as

efficiency and equity. Both concepts are crucial to the evaluation of the policy measure

from the research point of view. They will have rather little appeal on politicians and in

particular the concerned public. The distribution of benefits and costs is generally

considered more important politically than the size of the net benefits. Benefits and costs

may be either diffuse or concentrated, where the concentration of benefits and the diffusion

of costs are likely to have a positive effect on the acceptance of the measure when

compared to the reverse case. Still, individuals are more likely to oppose to new costs than

to support new benefits. Furthermore, even sophistically designed redistribution schemes

cannot resolve all equity concerns due to individual- and household-level variation in

flexibility and reaction potentials (alternatives) to the charging measure.

Concluding, the discussion about distributional effects and equity shows that the overall

“construct” of equity is very difficult to capture by the economic theory and quantitatively.

Since there does not exist an indisputable consensus about the definition of equity, the

concept (of an equitable distribution) lacks an applicable theoretical basis. Furthermore, it

remains questionable to what extent a universally equitable welfare situation can be

achieved through road charging revenue redistribution taking into account the plurality of

individual interests of car drivers and their unique perception of equity and fairness.

Nevertheless, due to the relevance of road charging as a “multifunctional” policy

instrument the evaluation of its economic impacts remains indispensable since it can

151

significantly help to develop a policy scheme that will be equitable enough to satisfy a

majority of the public involved in the measure implementation.

Finally, some ideas for future research in the field of this study shall be mentioned.

Depending on the policy concern to be examined using the modeling tool developed within

this work, the modeling mechanism can be further extended and refined. One possible idea

is the introduction of a bundle of strategies or policy instruments to reduce environmental

impacts of transport, besides road use charging and their simultaneous impact analysis. An

approach of this kind could help to better understand the effects from the implementation of

road charging and a simultaneous reduction in fuel and/ or other car related taxes.

Furthermore, the household income quartiles integrated in the model database could be

further disaggregated, e.g., as to income deciles. This would allow the implementation of

quantitative methods in the assessment of distributional and equity effects.

Another model extension could concern the combination of the travel demand data with

road infrastructure network data to allow for the modelling of road congestion. This kind of

model refinement could also include the introduction of household social utility functions

that would consider individual value of time and time use preferences.

The existence of valuable research ideas related to the work done within this dissertation

show – together with the results discussed above –the value of the research carried out here

and the benefit of the modelling tool developed within this work.

152

153

7 Bibliography

Abe, A. (1975) Distributional equity and the optimal pricing of urban transit. Journal of

Transport Economics and Policy, 9, 2, pp. 178-185.

AFFORD (2002) Schade J. and B. Schlag, Acceptability of Urban Transport Pricing.

Government Institute for Economic Research. Helsinki 2002.

http://www.vatt.fi/site/en/publications/latestPublications/publication/Publication_134

5_id/291 (24.11.2009).

Armington, P. (1969) A theory of demand for products distinguished by place of

production. IMF Staff Papers, 16, pp. 159-178.

Arnott, R., A. De Palma, and R. Lindsey (1994) The Welfare Effects of Congestion Tolls

with Heterogeneous Commuters. Journal of Transport Economics and Policy, 28, pp.

139-161.

Arrow, K. and G. Debreu (1954) The existence of an equilibrium for a competitive

economy, Econometrica, 22, pp. 265-290.

Atkinson, A. B. (1970) On the Measurement of Inequality. Journal of Economic Theory,

2(3), pp. 244-263.

Atkinson, A. B., L. Rainwater and T. M. Smeeding (1995) Income Distribution in OECD

Countries. OECD Social Policy Studies, No. 18, Paris.

Bach, S., Ch. Bork, V. Ehrenheim, M. Kohlhaas, K. Schumacher, Ch. Lutz, B. Meyer, B.

Praetorius, and H. Welsch (2001) Die oekologische Steuerreform in Deutschland:

Eine modellgestützte Analyse ihrer Wirkungen auf Wirtschaft und Umwelt.

Veroeffentlicht von Birkhäuser, 2001.

Bartley, B. (1995) Mobility impacts, reactions and opinions. Traffic engineering and

control, 36, pp. 596-603.

Bundesanstalt für Straßenwesen (BASt) (2005) Fahrleistungserhebung 2002, Berichte der

Bundesanstalt für Straßenwesen, Heft V 120, Bergisch-Gladbach.

Beckmann, M., C. B. McGuire, and Ch. B. Winsten (1956) Studies in the Economics of

Transportation. New Haven, Connecticut Yale, University Press.

Bell, M. G. H., M. A. Quddus, J.-D. Schmoecker, and A. M. Fonzone (2004) The Impact of

the Congestion Charge on the Retail Sector. Working Paper, London Centre for

Transport Studies, Imperial College.

154

Bergman, L. (1990) The Development of Computable General Equilibrim Modeling. In:

Bergman, L., D.W. Jorgenson, and E. Zalai (eds.): General Equilibrium Modeling

and Economic Policy Analysis, Cambridge, MA, pp. 3-30.

Bhattacharyya, S. C. (1996) Applied general equilibrium models for energy studies: a

survey. Energy Economics, Vol. 18, Issue 3, pp. 145-164.

Bibi S. and J. Duclos (2003) Equity and Policy Effectiveness with Imperfect Targeting.

Cahiers de recherche, 0335, CIRPEE.

Blanchard, W. (1986) Evaluating social equity: What does fairness mean and how can we

measure it? Policy Studies Journal 15(1), pp. 29-54.

Blum, U. C. H., G. Foos, and M. J. I. Gaudry (1988) Aggregate time series gasoline

demand models: Review of the literature and new evidence for West Germany.

Transportation Research Part A: General, Vol. 22, Issue 2, pp. 75-88.

Boehringer, C. and A. Loeschel (2006) Computable General Equilibrium Models for

Sustainability Impact Assessment: Status quo and Prospects. Ecological Economics

60(1), pp. 49-64.

Borges, A. (1986) Applied General Equilibrium Models: An Assessment of their

Usefulness for Policy Analysis. OECD Economic Studies, 7, pp. 7-43.

Bresson, G., J. Dargay, J.-L. Madre, and A. Pirotte (2004) Economic and structural

determinants of the demand for public transport: an analysis on a panel of French

urban areas using Shrinkage estimators. Transportation Research Part A: Policy and

Practice, Vol. 38, Issue 4, pp. 269-285.

Broecker, J. (2001) Spatial Effects of Transport Infrastructure: The Role of Market

Structure. In: Roy, J. and W. Schulz (eds), Theories of Regional Competition.

Nomos, Baden-Baden, pp. 181-193.

Broecker, J. (2002) Passenger Flows in CGE Models for Transport – Project Evaluation.

Institute for Regional Research Christian-Albrechts-Universitaet Kiel, Paper

presented at the ERSA Congress 2002, August 2002, Dortmund.

Broecker, J. (2004) Computable General Equilibrium Analysis in Transportation

Economics. In: Hensher, D. A., K. J. Button, K. E. Haynes, and P. R. Stopher (eds),

Handbook of Transport Geography and Spatial Systems. Elsevier, Amsterdam, 2004,

pp. 269-292.

Brooke, A., D. Kendrick, A. Meeraus, and R. Raman (1998) GAMS: A User’s Guide,

Washington D.C., GAMS Development Corp.

155

Bundesministerium für Verkehr, Bau- und Wohnungswesen (BMVBW) (ed.) (2002)

Analyse von Aenderungen des Mobilitätsverhaltens – insbesondere der Pkw-

Fahrleistung – als Reaktion auf geaenderte Kraftstoffpreise, Schlussbericht zum

Forschungsprojekt Nr. 96.0756/2002/, Bearb.: Institut für angewandte Verkehrs- und

Tourismusforschung e.V. (IVT) Heilbronn / Mannheim (Federführung), Progtrans

AG Basel, Steinbeis-Transferzentrum Angewandte Systemanalyse (STASA)

Stuttgart, Bonn.

Butler, M. and H. P. Williams (2002) Fairness versus efficiency in charging for the use of

common facilities. Journal of the Operational Research Society, 53(12), pp. 1324-

1329.

Button, K. (1993a) Transport Economics. Second edition. Edward Elgar Publishing,

Aldershot. Campbell, T. D..

Button, K. (1993b) Transport, the Environment and Economic Policy. Edward Elgar

Publishing, Cheltenham.

CAPRI (1999) Rennings, K., A. Ricci, and C. Sessa, Concerted Action on Transport

Pricing Research Integration (CAPRI), Deliverable D3 Valuation of transport

externalities. Brussels : European Commission, 1999.

CAPRI (2001) Nash, C., T. Sansom, and B. Matthews, Concerted Action on Transport

Pricing Research Integration (CAPRI), Final report for publication. Brussels :

European Commission, 2001.

CE Delft (2004), Climate Impacts from International Aviation and Shipping, Report for the

Netherlands Research Programme on Climate Change, Scientific Assessments and

Policy Analysis.

CE Delft (1999) Dings, J. M. W., P. Janse, B. A. Leurs, and M. D. Davidson, Efficient

Prices for transport, estimating the social costs of vehicle use. Delft: CE Delft, 1999.

CE Delft (2002a) Dings, J. M. W., R. C. N. Wit, B. A. Leurs, et al., External Costs of

Aviation. Delft: CE Delft, 2002.

CE Delft (2002) Dings, J. M. W., B.A. Leurs, A.F. Hof, D.M. Bakker, P.H. Mijer, E.T.

Verhoef, Returns on roads: Optimising road investments and use with the ‘user pays

principle’. CE Delft, 4Cast, Free University of Amsterdam Delft: CE Delft, 2002.

CE Delft (2003) Dings, J. M. W., M. D. Davidson, M.N. Sevenster, External and

Infrastructure costs of road and rail traffic: analysing European studies. Delft: CE

Delft, 2003.

156

CE Delft (2004a) Essen van, H. P., B. H. Boon, L. C. den Boer, J. Faber, M. A. van den

Bossche, K. T. H. Vervoort, and C. Rochez, Marginal costs of Infrastructure use –

towards a simplified approach. Delft: CE Delft, 2003.

CE Delft (2004b) Vermeulen, J. P. L., B. H. Boon, H. P. van Essen, L. C den Boer, J. M.

W. Dings, F. R. Bruinsma, M. J. Koetse, The price of Transport; overview of the

social costs of transport. Delft: CE Delft, 2004.

CE Delft (2008) Handbook on estimation of external costs in the transport sector:

Internalisation Measures and Policies for All external Costs of Transport, (IMPACT)

Delft.

Coase, R. H. (1946) The Marginal Cost Controversy. Economica, n.s. 13, August, pp. 169-

182. Reprinted as Chapter 4, pp. 75-93. In Coase’s The Firm, the Market, and the

Law. Chicago, University of Chicago Press, 1988.

Coase, R. H. (1960) The Problem of Social Cost, Journal of Law and Economics. October

1960, 3, pp. 1-44.

Coevering van de, P. and T. Schwanen (2006) Re-evaluating the impact of urban form on

travel patterns in Europe and North-America. Transport Policy, Vol. 13, Issue 3, pp.

229-239.

Cohen, Y. (1987) Commuter welfare under peak-period congestion tolls: Who gains and

who loses? International Journal of Transport Economics, 14, pp. 239-66.

Commission of the European Communities (CEC) (1995) Towards fair and efficient

pricing in transport, Policy options for internalising the external costs of transports in

the European Union, Green Paper, COM(95) 691 final. European Commission,

Brussels.

Commission of the European Communities (CEC) (2001) White Paper: European

Transport Policy for 2010: Time to Decide.

Conrad, K. (1999) Computable General Equilibrium Models for Environmental Economics

and Policy Analysis. In: van den Bergh, J.C.J.M. (ed.), Handbook of Environmental

and Resource Economics, Cheltenham, pp. 1061-1087.

Conrad, K. (2001) Computable General Equilibrium Models in Environmental and

Resource Economics. In: Tietenberg, T. and H. Folmer (eds.), The International

Yearbook of Environmental and Resource Economics 2002/2003, pp. 66-114.

CUPID (2003) Co-ordinating Urban Pricing Integrated Demonstrations. European

Commission Competitive and Sustainable Growth Programme. http://www.transport-

pricing.net/reports22.html (24.11.2009).

157

Dargay, J. (2001) The effect of income on car ownership: evidence of asymmetry.

Transportation Research Part A: Policy and Practice, Vol. 35, Issue 9, pp. 807-821.

Dargay, J. (2002) Determinants of car ownership in rural and urban areas: a pseudo-panel

analysis. Transportation Research Part E: Logistics and Transportation Review, Vol.

38, Issue 5, pp. 351-366.

Dargay, J. and D. Gately (1999) Income’s effect on car and vehicle ownership, worldwide:

1960-2015, Transportation Research Part A: Policy and Practice, Vol. 33, Issue 2, pp.

101-138.

Delucchi, M.A, (1997) The annualized social cost of motor-vehicle use in the US, based on

1990–1991 data: summary of theory, data, methods, and results. In: Greene, D.L.. D.

Jones, and M.A. Delucchi, (eds.) Full Costs and Benefits of Transportation. Springer-

Verlag, Berlin, Germany, pp. 27-68.

Delucchi, M. A. (2000) Environmental Externalities of Motor Vehicle Use. Journal of

Transport Economics and Policy, 34, pp. 135-168.

DESIRE (2001) Designs for Interurban Road pricing schemes in Europe. Project

Coordinator: TIS.PT, Consultores em Transportes, Inovação e Sistemas, S.A.

http://www.tis.pt/proj/Desire/6a.htm (24.11.2009).

Dupuit, J. (1844) De la Mesure de l'utilité des Travaux Publics. Annales des Ponts et

Chaussées: Mémoires et Documents, 2nd ser., 8(2), pp. 332-375. Trans. R.M.

Barback, On the Measurement of Utility for Public Works. International Economic

Papers, 1952, 2, pp. 83-110.

Dupuit, J. (1849) Des Péages. Annales des Ponts et Chaussées: Mémoires et Documents,

2nd ser., 17(1), pp. 207-248. Trans. Elizabeth Henderson, On Tolls and Transport

Charges. International Economic Papers, 1962, 11, pp. 7-31.

Eliasson, J. and L.-G. Mattsson (2006) Equity effects of congestion pricing: Quantitative

methodology and a case study for Stockholm. Transportation Research A 40, pp.602-

620.

Else, P. K. (1986) No Entry for Congestion Taxes? Transportation Research 30A 2, pp. 99-

107.

EURoPrice (2002) Consultation with the Business Community, Guidance Paper No. 2,

Bristol: EURoPrice Initiative Phase 2.

European Commission (EC) (1999) European Commission, 1999. Directive

1999/62/ECnext term of the European Parliament and of the Council of June, 17th,

previous term 1999 next term on the charging of heavy goods vehicles for the use of

158

certain infrastructures. Official Journal of the European Commission L187/42:

Brussels.

European Commission (EC) (2006) Directive 2006/38/EC amending Directive 1999/62/EC

on the charging of heavy goods vehicles for the use of certain infrastructures.

Brussels, 17 May 2006.

European Commission (EC) (2006), COM (2006) 314: Keep Europe Moving – Sustainable

mobility for out continent Midterm review of the European Commission’s 2001

White Paper, Brussels

European Commission (EC) (2008) Proposal for a directive of the European Parliament and

of the council amending Directive 1999/62/EC on the charging of heavy goods

vehicles for the use of certain infrastructures.

European Commission (EC) (2009) European Parliament legislative resolution of 11 March

2009 on the proposal for a directive of the European Parliament and of the Council

amending Directive 1999/62/EC on the charging of heavy goods vehicles for the use

of certain infrastructures.

Eurostat (2008) Eurostat Manual of Supply, Use and Input-Output Tables. Luxembourg:

Office for Official Publications of the European Communities.

Evans, A.W. (1992) Road congestion pricing: When is it a good policy? Journal of

Transport Economics and Policy, 26 (September), pp.213-243.

ExternE (1996) Externalities of Energy, Volume 2: Methodology, Luxembourg: European

Commission, 1996.

ExternE (1997) P. Bickel (IER) et al. External Costs of Transport in ExternE, Luxembourg:

European Commission, 1997.

ExternE (1998) N. Eyre et al. Global Warming Damages. Luxembourg: European

Commission, 1998.

ExternE (1999a) Externalalities of Energy: Methodology Annexes. Luxembourg: European

Commission, 1999.

ExternE (1999b) Externalities of Energy, Volume 7: Methodology 1998 update, 1999.

Luxembourg: European Commission, 1999.

ExternE (1999c) External Costs of Energy Conversion: Improvement of the ExternE

Methodology and Assessment of Energy-related Transport Externalities, Final Report

Stuttgart: IER Stuttgart, 1999.

159

ExternE (2005) P. Bickel, R. Friedrich (eds.) Externalities of Energy, Methodology 2005

update. Luxembourg: European Commission, 2005.

Farrell, S. and W. Saleh (2005) Road-user charging and the modelling of revenue

allocation. Transport Policy, Vol. 12, Issue 5, pp. 431-442.

Fehr, H. and W. Wiegard (1996) Numerische Gleichgewichtsmodelle: Grundstruktur,

Anwendungenund Erkenntnisgehalt. In: Experimente in der Oekonomie, Oekonomie

und Gesellschaft, Jahrbuch 13, Campus-Verlag, Frankfurt, pp. 296-339.

Feldstein, M. and H. S. Luft (1973) Distributional Constraints in Public Expenditure

Planning. Management Sciences 19, 12 (August 1973), pp. 1414-1422.

Franklin, J. P. (2004), Non-parametric distributional analysis of a transportation policy: the

case of Stockholm’s congestion pricing trial. Paper presented at the Transportation

Research Board 84th Annual Meeting, Washington, DC, January 2005.

Fridström, L. (2000) Vinnare och förlorare i olika typer av avgiftssystem:

modellberäkningar för Oslo. TÖI arbetspapper. Del av AFFORD projektet.

Gaunt, M., T. Rye, and S. Allen (2007) Public acceptability of road user charging: the case

of Edinburgh and the 2005 Referendum. Transport Reviews 27 (1).

German Federal Office for Building and Regional Planning (BBSR) (2005),

http://www.bbr.bund.de/cln_005/nn_117070/DE/Raumbeobachtung/Werkzeuge/Rau

mabgrenzungen/raumabgrenzungen__node.html?__nnn=true ,

http://www.bbr.bund.de/cln_005/nn_103086/DE/Raumbeobachtung/Werkzeuge/Rau

mabgrenzungen/SiedlungsstrukturelleGebietstypen/Regionstypen/regionstypen.html

(03.03.2009).

Ginsburgh V. and M. Keyzer (1997) The Structure of Applied General Equilibrium

Models. MIT Press, Cambridge, Massachusetts.

Giuliano, G. (1992) An Assessment of the Political Acceptability of Congestion Pricing.

Transportation, 19, 4, pp.335-58.

Giuliano, G. (1994) Equity and Fairness Considerations of Congestion Pricing. In:

Committee for Study on Urban Transportation Congestion Pricing (ed.), Curbing

Gridlock: Peak Period Fees to Relieve Traffic Congestion. Transportation Research

Board Special Report 242, Vol. 2, National Academy Press. Washington, D.C., pp.

250-279.

Giuliano, G. and J. Dargay (2006) Car ownership, travel and land use: a comparison of the

US and Great Britain. Transportation Research Part A: Policy and Practice, Vol. 40,

Issue 2, pp. 106-124.

160

Glazer, A., H. Link, A. May, D. Milne, and E. Niskanen (2001) Barriers to transport

pricing– a review of research, Presented to IMPRINT-EUROPE Seminar, Brussels.

Glazer, A. and E. Niskanen (2000) Which consumers benefit from congestion tolls? Journal

of Transport Economics and Policy 34 (January), pp. 43-54.

Goodwin, P. (1992) A review of new demand elasticities with reference to short and long

run effects of price changes. Journal of Transport Economies and Policy, 26, pp. 155-

169.

Goodwin, P. (1997) Solving Congestion: Inaugural Lecture of the Professorship of

Transport Policy, University College London,

http://www.cts.ucl.ac.uk/tsu/pbginau.htm (24.11.2009).

Goodwin, P., J. Dargay, and M. Hanly (2004) Elasticities of Road Traffic and Fuel

Consumption with Respect to Price and Income: A Review, Transport Reviews, Vol.

24, No. 3, pp. 275-292.

Gottfried, P.E. Stöß and W. Wiegard (1990) Applied General Equilibrium Tax Models:

Prosepcts, Examples, Limits. In: Petersen H.-G. and J. K. Brunner (eds.), Simulation

Models in Tax and Transfer Policy, Frankfurt, pp. 205-344.

GRACE ITS, University of Leeds (Project coordinator) GRACE: Generalisation of

research on accounts and cost estimation). On-going Leeds: ITS University of Leeds.

GRACE (2005) Link, H. et al., Generalisation of Research on Accounts and Cost

Estimation (GRACE). Deliverable 1: Information Requirements for Monitoring

Implementation of Social Marginal Cost Pricing. Leeds: ITS, University of Leeds,

2005.

GRACE (2006a) G. Lindberg, Generalisation of Research on Accounts and Cost

Estimation (GRACE) Deliverable D3 Marginal Costs Case Studies for Road and Rail

Transport. Leeds: ITS, University of Leeds, 2006.

GRACE (2006b) G. Lindberg, Generalisastion of Research on Accounts and Cost

Estimation (GRACE). Case study 1.4: State-of-the-art of external marginal accident

costs, Annex to Deliverable D3 Marginal Coast Case Studies for Road and Rail

Transport. Leeds: ITS, University of Leeds, 2006.

GRACE (2006c) Lieb, C., S. Suter, and P. Bickel, Generalisation of Research on Accounts

and Cost Estimation (GRACE) Environmental Costs in Sensitive Areas. Leeds: ITS,

University of Leeds, August 2006.

GRACE (2006d) Meersman, H., F, Monteiro, T. Pauwels, E. Van de Voorde, T.

Vanelslander, et al., Generalisation of Research on Accounts and Cost Estimation

161

(GRACE) Deliverable D4: Marginal Cost Case Studies for Air and Water Transport.

Antwerp: University of Antwerp, 2006.

Graham, D. and S. Glaister (2002a) The Demand for Automobile Fuel: A Survey of

Elasticities. Journal of Transport Economics and Policy, 36, pp. 1-26.

Graham, D. and S. Glaister (2002b) Review of Income and Price Elasticities in the Demand

for Road Traffic. London, Department for Transport.

Graham, D. and S. Glaister (2004) Road Traffic Demand Elasticity Estimates: A Review,

Transport Reviews, Vol. 24, No. 3, pp. 261-274.

Gray, D. and D. Begg (2001) Delivering congestion charging in the UK: what is required

for its successful introduction? Policy Paper Series No.4. The Centre for Transport

Policy. The Robert Gordon University, Aberdeen.

Gunning, J.W. und M.A. Keyzer (1995), Applied General Equilibrium Models for Policy

Analysis. In: Handbook of Development Economics, Vol. 3A, pp. 2025-2107.

Hanly, M., J. Dargay and P. Goodwin (2002) Review of Income and Price Elasticities in

the Demand for Road Traffic. London: Department for Transport.

Harrison, E. and C. Seidl C. (1994) Acceptance of distributional axioms: Experimental

findings. In Eichhorn, W. (ed.), Models and measurement of welfare and inequality,

Springer-Verlag, pp.67-94.

Harrington, W., A. J. Krupnick, and A. Alberini (2001) Overcoming public aversion to

congestion pricing. Transportation Research Part A, 35, pp. 87-105.

Hau, T.D. (1992) Economic Fundamentals of Road Pricing. Policy Research Working

Paper, Transport Division, Infrastructure and Urban Development Department, The

World Bank, Washington, DC.

Hau, T. D. (1998) Congestion Pricing and Road Investment. In Road Pricing, Traffic

Congestion and the Environment: Issues of Efficiency and Social Feasibility. Button,

K. J. and E. T. Verhoef (eds.), Cheltenham, UK, Edward Elgar, pp. 39-78.

Hau, T. D. (2005a) Economic Fundamentals of Road Pricing: A Diagrammatic Analysis,

Part I-Fundamentals. Transportmetrica 1(2), pp. 81-115.

Hau, T. D. (2005b.) Economic Fundamentals of Road Pricing: A Diagrammatic Analysis,

Part II-Relaxation of Assumptions. Transportmetrica 1(2), pp. 119-149.

Hautzinger, H. (1978) Travel demand elasticity with respect to socio-economic variables:

The multinomial logit case. Transportation Research, Vol. 12, Issue 5, pp. 355-356.

162

Herry, M. and N. Sedlacek (2003) Österreichische Wegekostenrechnung für die Straße

2000. Austrian Ministry of Transport, Innovation and Technology, Vienna.

Hicks, J. R. (1939) Value and Capital: An Inquiry into Some Fundamental Principles of

Economic Theory. Oxford, Clarendon Press.

Hochschild, J. L. (1981) What's Fair? Harvard University Press, Cambridge, Mass.

IMPACT (2008) Handbook on estimation of external costs in the transport sector. Produced

within the study Internalisation Measures and Policies for All external Cost of

Transport (IMPACT).

http://ec.europa.eu/transport/sustainable/studies/sustainable_en.htm (24.11.2009).

IMPRINT (2004) Matthews, B. and C. Nash, Implementing Pricing Reform in Transport –

Effective Use of Research on Pricing in Europe. Project Partners: Institute for

Transport Studies (ITS), Leeds, Forschungs- und Anwendungsverbund

Verkehrssystemtechnik (FAV), Berlin, Istituto di Studi per l'Integrazione dei Sistemi

(ISIS), Rome, Netherlands Organization for Applied Scientific Research (TNO),

Delft, The Government Institute for Economic Research (VATT), Helsinki, Technical

and Economical University of Budapest, Department of Transport Economics

(BUTE), Budapest. http://www.imprint-eu.org/ (24.11.2009).

Infras/ IWW (2000) Maibach, M. et al., External Costs of Transport, Infras, IWW, im

Auftrag der International Union of Railways (UIC), Paris.

Infras /IWW (2004) External Costs of Transport – Update Study, im Auftrag der

International Union of Railways (UIC), Final Report, Zuerich/ Karlsruhe.

IVT Institut für angewandte Verkehrs- und Tourismusforschung, progtrans, stasa (2004):

Analyse von Änderungen des Mobilitätsverhaltens - insbesondere der Pkw-

Fahrleistung - als Re-aktion auf geänderte Kraftstoffpreise, Schlussbericht zum

Forschungsprojekt Nr. 96.0756/2002 des Bundesministeriums für Verkehr, Bau- und

Wohnungswesen.

Ison, S. (2004) Road User Charging: Issues and Policies. Hampshire, UK: Ashgate

Publishing Ltd.

Jaensirisak, S., A. D. May, and M. R. Wardman (2003) Acceptability of road user charging:

the influence of selfish and social perspectives. In J. Schade and B. Schlag (eds.),

Acceptability of transport pricing strategies. Oxford, Elsevier, pp. 203-218.

Jaensirisak, S., M. Wardman, and A. D. May (2005) Explaining variations in public

acceptability of road pricing schemes. Journal of Transport Economics and Policy,

39, 2, pp. 127-153.

163

Jansson, J. O. (1997) Theory and practice of transport infrastructure and public transport

pricing. In: de Rus, G. and C. Nash (eds.) 1997. Recent Developments in Transport

Economics. Ashgate, Aldershot, UK.

Jha, R (1998) Modern public economics. Routledge, London and New York.

Johansson, M. V., T. Heldt, and P. Johansson (2006) The effects of attitudes and

personality traits on mode choice. Transportation Research Part A: Policy and

Practice, Vol. 40, Issue 6, pp. 507-525.

Johansson, O. and L. Schipper (1997) Measuring the long run fuel demand of cars: separate

estimations of vehicle stock, mean fuel intensity, and mean annual driving distance.

Journal of Transport Economics and Policy, 31, pp. 277-292.

Johansson-Stenman, O. (2002) Estimating Individual Driving Distance by Car and Public

Transport Use in Sweden. Applied Economics. Taylor and Francis Journals, Vol.

34(8) May, pp. 959-67.

Jones, P. M. (1991) Gaining public support for road pricing through a package approach.

Traffic engineering and control, 32, pp. 194-196.

Jones, P. (1992) Some fundamental issues underlying value of time research and its

application to project evaluation. In: Notes prepared for a seminar on value of travel

time at the Institute of Transport Economics. Oslo.

Jones, P. (1998) Urban Road Pricing: Public Acceptability and Barriers to Implementation.

In Button, K. J. and E. T. Verhoef (eds.), Road Pricing, Traffic Congestion and the

Environment, Edward Elgar, USA.

Jones, P. (2003) Acceptability of road user charging: Meeting the challenge. In J. Schade J.

and B. Schlag (eds.), Acceptability of transport pricing strategies, Oxford, Elsevier,

pp. 27-62.

Just, R. E., D. L. Hueth, and A. Schmitz (2004) The Welfare Economics of Public Policy:

A Practical Approach to Project and Policy Evaluation. Edward Elgar, Cheltenham.

Kalinowska D. and H. Kuhfeld (2006) Motor Vehicle Use and Travel Behaviour in

Germany – Determinants of Car Mileage, DIW Berlin Discussion Papers 602, Berlin.

Kalinowska, D., J. Kloas, H. Kuhfeld, and U. Kunert (2005) Aktualisierung und

Weiterentwicklung der Berechnungsmodelle für die Fahrleistung von

Kraftfahrzeugen und für das Aufkommen und für die Verkehrsleistung im

Personenverkehr (MIV), Gutachten im Auftrag des Bundesministeriums für Verkehr,

Bau- und Wohnungswesen. Berlin, Projektnummer 19.0005/2002.

164

Kalmanje, S. and K. M. Kockelman (2004) Credit-based congestion pricing: travel, land

value and welfare impacts. Transportation Research Record 1864, pp. 45-53.

Kehoe, P. J. and T. J. Kehoe (1994) A primer on Static Applied General Equilibrium

Models. Ferderal Reserve Bank of Minneapolis Quarterly review, Spring 1994,

Volume 18, 1.

Kelly, N. J. (2005) Political Choice, Public Policy, and Distributional Outcomes. American

Journal of Political Science, Vol. 49, No. 4 (Oct., 2005), pp. 865-880.

Kiefer, D. W. (1983) Measurement of the Progressivity of Public Expenditures and Net

Fiscal Incidence: Comment. Southern Economic Journal, Vol. 50, No. 2 (Oct., 1983),

pp. 578-586. Published by: Southern Economic Association.

King, S., and S. M. Sheffrin (2002) Tax evasion and equity theory: An investigative

approach. International Tax and Public Finance, 9, pp. 505-521.

Klepper, G., J.-O. Lorz, F. Staehler, R. Thiele and M. Wiebelt (1994) Empirische

allgemeine Gleichgewichtsmodelle – Struktur und Anwendungsmöglichkeiten.

Jahrbuecher fuer Nationaloekonomie und Statistik, 213, pp. 513-544.

Kletzan, D., A. Köppl, K. Kratena, S. Schleicher, and M. Wüger (2006) Towards

sustainable consumption: Economic modelling of mobility and heating for Austria,

Ecological Economics, Volume 57, Issue 4, 1 June 2006, pp. 608-626.

Knight, F. (1924) Some fallacies in the interpretation of social cost, Quarterly Journal of

Economics 38 (1924), pp. 582-606.

Kunert, U. and R. Follmer (2005) Methodological Advances in National Travel Surveys:

Mobility in Germany 2002. Transport Reviews, 25,4, pp.415-431.

Langmyhr, T. (1995) Planning for road pricing: A multi-rationality approach. Scandinavian

Housing and Planning Research 12 2, pp. 73-91.

Langmyhr, T. (1997) Managing Equity: The case of road pricing in Transport Policy, Vol.

4, No. 1, pp. 25-39.

Lee, D. B. (1993) Full Costs of Pricing Highways. Department of Transportation,

Washington, DC.

Leontief, W. (1941) The Structure of American Economy, 1919-1939: An empirical

application of equilibrium analysis. 1951 edition, New York, Oxford University

Press.

Levy, F. S., A. J. Meltsner, and A. Wildavsky (1974), Urban Outcomes, University of

California Press, Berkeley.

165

Limtanakool, N., M. Dijst, and T. Schwanen (2006) The influence of socioeconomic

characteristics, land use and travel time considerations on mode choice for medium-

and longer-distance trips. Journal of Transport Geography, Vol. 14, Issue 5, pp. 327-

341.

Lindsey, C.R. and E.T. Verhoef, (2001) Traffic congestion and congestion pricing. In:

Hensher, D.A., and K.J. Button (eds.), Handbook of Transport Systems and Traffic

Control. Handbooks in Transport 3. Elsevier, Amsterdam, pp. 77-105.

Lindsey, R. (2003) Road pricing issues and experiences in the US and Canada. Imprint-

Europe 4th Seminar Implementing Pricing Policies in Transport. Phasingand

Packaging, Katholieke University of Leuven, Belgium.

Link, H. and J. Polak (2001) How acceptable are transport pricing measures? Empirical

studies in nine European Countries. Proceedings of the 29th European Transport

Conference, PTRC, London.

Lipsey, R.G. and K.J. Lancaster (1956) The general theory of second best. Review of

Economics Studies, Vol. 24, pp.11-32.

Litman, T. (2003) Transportation Cost and Benefit Analysis: Techniques, Estimates and

Implications. Victoria Transport Policy Institute, Victoria, Canada. www.vtpi.org/tca

(24.11.2009).

Lucy, W. (1981) Equity and planning for local services, Journal of the American Planning

Association, Journal of the American Planning Association, Vol. 47, Issue 4, pp. 447-

457.

Luk, J. and E. Chung. (1997) Induced demand and road investment: An initial appraisal.

Research Report, ARR 299. Vermont South, Australia: ARRB Transport Research.

Mansur, A. and J. Whalley (1984) Numerical Specification of Applied General Equilibrium

Models: Estimation, Calibration and Data. In Scarf H. and J. B. Shoven (eds.), pp. 69-

126.

Marshall, A. (1890) Principles of Economics. Macmillan, London.

Mayeres, I. (1998) The Distributional Impacts of Policies for the Control of Transport

Externalities – An Applied General Equilibrium Model, Center for Economic Studies

– K. U. Leuven.

Mayeres, I. (2000) The Efficiency Effects of Transport Policies in the Presence of

Externalities and Distortionary Taxes. Journal of Transport Economics and Policy,

Vol. 34, Part 2, pp. 233-260.

166

Mayeres, I. (2001) Equity and Transport Policy Reform. ETE Discussion Paper 2001-14,

CES, K. U. Leuven, Leuven.

Mayeres, I. (2004) Testing alternative transport pricing strategies: A CGE analysis for

Belgium. Paper presented at the Conference on Input-Output and General

Equilibrium: Data, Modelling and Policy Analysis, Brussels, 2-4 September 2004.

Mayeres, I. and S. Proost (2002) Reforming transport pricing: an economist's perspective

on equity, efficiency and acceptability. Energy, Transport and Environment Working

Papers Series ete0212. Katholieke Universiteit Leuven, Centrum voor Economische

Studien, Energy, Transport and Environment.

MC-ICAM (2004) Marginal Cost Pricing in Transport – Integrated Conceptual and Applied

Model Analysis. Project Coordinator: The Institute for Transport Studies, University

of Leeds. http://www.its.leeds.ac.uk/projects/mcicam/index.html (24.11.2009).

Meyer, B. and G. Ewerhart (1998) Multisectoral Policy Modelling for Environmental

Analysis. In: Uno, K. / Bartelmus, P. (eds.), Environmental Accounting in Theory

and Practice. Amsterdam, pp. 395-406.

MiD (2002) infas and DIW Berlin, Survey Mobilität in Deutschland.

http://www.mid2002.de/engl/index.htm, (12.10.2006).

Miller, D. (1976) Social Justice. Oxford, Oxford University Press.

Mohring, H. (1964) Relation Between Optimum Congestion Tolls and Present Highway

User Charges. Traffic Congestion as a Factor in Road-User Taxation, Highway

Research Record 47, pp. 1-14.

Mohring, H. (1976) Transportation Economics. Ballinger, Cambridge, Mass.

Mohring, H. and M. Harwitz (1962) Highway Benefits: An Analytical Framework.

Evanston, Il, Northwestern University Press.

Morrison, S. (1986) A Survey of Road Pricing. Transportation Research, 20 A, pp. 87-97.

Muellbauer, J. (1973) Prices and Inequality: The Recent U.K. Experience, Discussion Paper

No. 6, University London, Birkbeck College.

Muellbauer, J. (1974a) Household composition, Engel curves and welfare comparisons

between households: A duality approach. European Economic Review 5, pp. 103-

122.

Muellbauer, J. (1974b) Inequality Measures, Prices and Household Composition. Review

of Economic Studies, pp. 493-502.

167

Munk, J. K. (2003) Assessment of the introduction of road pricing using a Computable

General Equilibrium mode. Institute of Economics, University of Aarhus.

Murphy, J. J. and M. A. Delucchi (1998) A Review of the literature on the Social Cost of

Motor-Vehicle Use in the United States. Journal of Transportation and Statistics, 1,

pp. 15-42.

Naess, P. (2006) Accessibility, Activity Participation and Location of Activities: Exploring

the Links between Residential Location and Travel Behavior, Urban Studies, Vol. 43,

No. 3, pp. 627-652.

Nash, C. (2001) Equity versus Efficiency in Transport Systems. In: Button, K. and D.A.

Hensher (eds.). Handbook of Transport Systems and Traffic Control, Pergamon,

Amsterdam, pp.33-45.

Newbery D. (1988) Road user charges in Britain. Economic Journal, 98 (Conference issue),

pp. 161-76.

Newbery, D. (1989) Cost Recovery from Optimally Designed Roads. Economica, 56, pp.

165-85.

Oberholzer-Gee, F. and H. Weck-Hannemann (2002) Pricing road use: politico-economic

and fairness considerations. Transportation Research D 7, pp. 357-371.

Ochmann, R., and A. Peichl (2006) Measuring Distributional Effects of Fiscal Reforms.

Finanzwissenschaftliche Diskussionsbeiträge, 06-9, Köln.

OECD (1982) The OECD List of Social Indicators, Paris.

OECD (2008) Growing Unequal? Income Distribution and Poverty in OECD Countries.

Paris. www.oecd.org/els/social/inequality / www.oecd.org/els/social/inegalite

(24.11.2009).

Owens, S. (1995) Predict and provide or predict and prevent?: pricing and planning in

transport policy. Transport Policy 2 1, pp. 43-49.

Paltsev, S. V. (1999) Moving from Static to Dynamic General Equilibrium Economic

Models: Notes for a beginner in MPSGE, Colorado, 1999.

Parry, M.L., O.F. Canziani, J.P. Palutik et al. (2007) Technical summary. In: M.L. Parry,

O.F. Canziani and J.P. Palutikof et al. (eds.), Climate Change 2007: Impacts,

Adaptation and Vulnerability, CUP, Cambridge, UK (2007), pp. 23-78.

PATS Consortium (2001) Recommendations on Transport Pricing Strategies, Final Report

of the PATS Project, (Brussels: European Commission).

168

Paulley N. (2002) Recent studies on key issues in road pricing, Transport Policy, 7, pp.

175-177.

Peireira, A. M. and J. B. Shoven (1988) Survey of dynamic computational general

equilibrium models for tax policy evaluation. Journal of Policy Modeling 10, pp.401-

426.

Piggot, J. and J. Whalley (1985) New Developments in Applied General Equilibrium

Analysis, Cambridge.

Piggot, J. and J. Whalley (1991) Applied General Equilibrium, Springer-Verlag, Berlin.

Pigou, A. C. (1920) The Economics of Welfare. Macmillan, London.

PRIMA (2000) Deliverable D4. Ways and Means to Improve the Acceptance of Urban

Road Pricing. Pricing Measures Acceptance Project, Inregia, Stockholm.

PROGRESS (2004) Competitive and Sustainable Growth Programme PROGRESS Project

2000-CM.10390. Pricing Road Use for Greater Responsibility, Efficiency and

Sustainability in Cities. http://www.progress-project.org/ (24.11.2009).

Proost, S. and K. Van Dender (2003) Marginal Social Cost Pricing for All Transport Modes

and the Effects of Modal Budget Constraints. Transportation Economics, Elsevier,

Vol. 9(1), pp. 159-177.

Pucher, J. and J. Renne (2003) Socioeconomics of Urban Travel: Evidence From the 2001

NHTS. Transportation Quarterly 57 (2003), pp. 49-77.

Quinet, E. (2004) A meta-analysis of Western European external costs estimates.

Transportation Research D, 9, pp. 465-476.

Ramsey, F. P. (1928) A mathematical theory of saving. Economic Journal, 38, pp. 548-559.

Raux, C., and S. S. Souche (2004) The Acceptability of Urban Road Pricing: A Theoretical

Analysis Applied to Experience in Lyon. Journal of Transport Economics and Policy,

Vol. 38, No. 2, pp. 191-216.

Rawls, J. (1971) A Theory of Justice. Harvard University Press, Cambridge, MA.

Rawls, J. (1999) A Theory of Justice. Revised. Belknap Press, Cambridge, MA.

RECORDIT (2003) Real Cost Reduction of Door-to-door Intermodal Transport

(RECORDIT). http://www.recordit.org/ (10/03/2008).

Richardson, H. W. (1974) A note on the distributional effects of road pricing. Journal of

Transport Economics and Policy, 8, pp. 82-85.

169

Richardson, H. W. and C. B. Chang-Hee (1998) The Equity Impacts of Road Congestion

Pricing. In Button K. J. and E. Verhoef (eds.), Road Pricing, Traffic Congestion and

the Environment. Issues of Efficiency and Social Feasibility. Edward Elgar,

Cheltenham, pp. 247-262.

Robinson, S. (1999), From Stylized to Applied Models: Building Multisector CGE Models

for Policy Analysis, North American Journal of Economics and Finance, S. 5-38.

Rothengatter, W. (2003) How good is first best? Marginal cost and other pricing principles

for user charging in transport. Transport Policy 10, 2, pp. 121-130.

Rouwendal, J. and E. Verhoef (2006) Basic economic principles of road pricing: from

theory to applications. Transport Policy, Vol. 13, Issue 2, pp. 106-114.

Rutherford, T. (1999) Applied General Equilibrium Modeling with MPSGE as a GAMS

Subsystem: An Overview of the Modeling Framework and Syntax. Computational

Economics, 14, pp. 1-46.

Ryan, W. (1981) Equality. New York, Pantheon Books.

Safirova, E., W. Harrington, P. Nelson, I. W. Parry, K. Gillingham, and D. Mason (2003)

Welfare and Distributional Effects of Road Pricing Schemes for Metropolitan

Washington, DC. In Santos, G. (ed.), Road Pricing: Theory and Evidence, Elsevier

Science.

Safirova, E., W. Harrington, P. Nelson, and D. Mason (2002) Distributional Consequences

of Road Pricing Policies on a Network Representing a Real City. Presented at the

World Congress of Environmental and Resource Economists (WCEE) in Monterey,

CA, June 2002.

Santos, G. and L. Rojey (2004) Distributional Impacts of Road Pricing: The Truth Behind

the Myth, Transportation, 31, pp. 21-42.

Schade, J. and B. Schlag (2000a) Acceptability of Urban Transport Pricing. Research

Report 72, Government Institute for Economic Research, Helsinki.

Schade, J. and B. Schlag (2000b) Public Acceptability of Traffic Demand Management and

Pricing Measures in Europe, Traffic Engineering and Control, 41, pp. 314-8.

Schade, J. and B. Schlag (2003) Acceptability of pricing reform. Paper presented to

IMPRINT-Europe Seminar.

Schaefer, A. and H. D. Jacoby (2005) Technology detail in a multisector CGE model:

transport under climate policy. Energy Economics, Vol. 27, Issue 1, pp. 1-24.

170

Schaefer, A. and H. D. Jacoby (2006) Vehicle technology under CO

constraint: a general

equilibrium analysis, Energy Policy, Vol. 34, Issue 9, pp. 975-985.

Schiller, E. (1998) The Road Ahead – the Economic and Environmental Benefits of

Congestion Pricing, Pacific Research Organization.

www.pacificresearch.org/issues/enviro/ congestion.htm (24.11.2009).

Schuert, A., M. Spangenberg, and T. Pütz (2005) Raumstrukturtypen: Konzept –

Ergebnisse – Anwendungsmöglichkeiten – Perspektiven, BBR-Arbeitspapier, Bonn,

31.05.2005,

http://www.bbr.bund.de/cln_007/nn_21360/DE/Raumbeobachtung/Werkzeuge/Raum

abgrenzungen/Raumstrukturtypen/Methodik/MethodikRstrukturTypen.html

(03.03.2009).

Shoven, J. B. and J. Whalley (1992) Applying General Equilibrium. Cambridge Surveys of

Economic Literature. Cambridge, University Press.

Shoven, J. B. and J. Whalley (1984) Applied General-Equilibrium Models of Taxation and

International Trade: An Introduction and Survey, Journal of Economic Literature,

American Economic Association, Vol. 22(3), pp. 1007-51.

Sikow-Magny, C. (2003) Efficient pricing in transport – overview of European ommission's

transport research programme. In: Schade, J. and B. Schlag (eds.), Acceptability of

Transport Pricing Strategies. Elsevier, Amsterdam (2003), pp. 13-26.

Small, K. A. (1983) The Incidence of Congestion Tolls on Urban Highways. Journal of

Urban Economics, 13, pp. 90-111.

Small, K. A. (1992) Using the Revenues of Congestion Pricing. Transportation, Vol. 19,

No. 4, pp. 359-381.

Small K. A. (1992) Special issue: Congestion pricing. Transportation, 19, pp. 287-291.

Small, K. A. and C. Kazimi (1995) On the Costs of Air Pollution from Motor Vehicles.

Journal of Transport Economics and Policy, 29(1), pp. 7-32.

Small, K. A. and C. Winston (1988) Optimal Highway Durability. American Economic

Review, June 1988, 78, pp. 560-569.

Small, K. A., C. Winston, and C. A. Evans (1989) Road Work: A New Highway Pricing

and Investment Policy. Washington, DC: The Brookings Institution.

Smith, A. (1937) An Inquiry into the Nature and Causes of the Wealth of Nations, Modern

Library edition. Edwin Cannan (ed.) New York: The Modern Library.

171

StaBuA, Statistisches Bundesamt (2004) Input-Output-Rechnung 2000,

Volkswirtschaftliche Gesamtrechnungen, Fachserie 18, Reihe 2, Wiesbaden.

StaBuA, Statistisches Bundesamt (2005a) Einkommens- und Verbrauchsstichprobe -

Einnahmen und Ausgaben Privater Haushalte 2003, Wirtschaftsrechnungen,

Fachserie 15, Heft 4, Wiesbaden.

StaBuA, Statistisches Bundesamt (2005b) Einkommens- und Verbrauchsstichprobe -

Aufwendungen privater Haushalte für den Privaten Konsum 2003,

Wirtschaftsrechnungen, Fachserie 15, Heft 5, Wiesbaden.

StaBuA, Statistisches Bundesamt (2005c) Einnahmen und Ausgaben Privater Haushalte

2003, Wirtschaftsrechnungen (LWR), Fachserie 15, Reihe 1, Wiesbaden.

StaBuA, Statistisches Bundesamt (2006a) Einkommens- und Verbrauchsstichprobe 2003,

Aufgabe, Methode und Durchführung der EVS 2003, Wiesbaden, Metzler-Poeschel,

Wirtschaftsrechnungen, Fachserie 15, Heft 7, Wiesbaden.

StaBuA, Statistisches Bundesamt (2006b) Inlandsproduktsberechnung: Detaillierte

Jahresergebnisse 2005, Volkswirtschaftliche Gesamtrechnungen, Fachserie 18, Reihe

1.4, Wiesbaden.

Stata, Data Analysis and Statistical Software. www.stata.com (24.11.2009).

Steininger, K. (2002) Environmentally Counterproductive Support Measures in Transport:

A CGE Analysis for Austria, University of Graz, Department of Economics,

Submitted to the 2002 World Congress of Environmental and Resource Economists.

Steininger, K. and B. Friedl (2004) Economic and Distributional Impacts of Nationwide

Car Road Pricing: A CGE Analysis for Austria. Paper submitted to the Thirteenth

Annual Conference of the European Association of Environmental and Resource

Economists, Budapest, June 2004.

Steininger, K., B. Friedl, and B. Gebetsroither (2007) Sustainability impacts of car road

pricing: A computable general equilibrium analysis for Austria. Ecological

Economics, Volume 63, Issue 1, 15 June 2007, pp. 59-69.

Stone, R. (1949) Social Accounting, Aggregation and Invariance. In Cahiers du Congrès

International de Comptabilité, 1948, French translation: EconomieAppliquée, Vol. II,

No. 1, pp. 26-54.

Stone, R. (1951) Simple Transaction Models, Information and Computing, paper presented

at a conference on Automatic Control. Cranfield, 1951, in The Review of Economic

Studies, Vol. XIX, no. 2, pp. 67 –84.

172

Stone, Richard (1955a), Model-Building and the Social Accounts: A Survey, in

International Association for Research in Income and Wealth (ed.), Income and

Wealth, Series IV, Bowes and Bowes, Cambridge.

Stone, R. (1955b) Input-Output and the Social Accounts, in University of Pisa (ed.), The

Structural Interdependence of the Economy. Proceedings of an International

Conference on Input-Output Analysis, Varenna, J. Wiley, New York, Giuffre, Milan,

pp. 155 -172.

Stone, R. and G. Croft-Murray (1959) Social Accounting and Economic Models, London.

Suits, D. B. (1977) Measurement of Tax Progressivity. The American Economic Review,

Vol. 67, No. 4 (Sep., 1977), pp. 747-752. Published by: American Economic

Association.

Thurman, W. and M. K. Wohlgenant (1989) Consistent Estimation of General Equilibrium

Welfare Effects. American Journal of Agricultural Economics, 71 (Nov.), pp. 1041-

45.

Transek, A. B. (2002) Road Pricing in Urban Areas. The Federation of European Transport

and Environment and the Swedish National Road Administration (Vagverket).

TRENEN-II-STRAN (1999) Proost, S. and K. van Dender, TRENEN II STRAN - Final

Report for Publication. Study funded by the 4th RTD framework programme of the

European Commission. Leuven : Katholieke Universiteit, Centre for Economic

Studies, 1999.

Tsekeris, T. and S. Voss (2009) Design and evaluation of road pricing: state-of-the-art and

methodological advances. Netnomics, Volume 10, Number 1 / April 2009, pp. 5-52.

Springer Netherlands.

U.K. Ministry of Transport. (1964) Road Pricing: The Economic and Technical

Possibilities. The Smeed Report. London: H.M. Stationery Office.

UNITE (2002) (Unification of accounts and marginal costs for Transport Efficiency

UNITE, http://www.its.leeds.ac.uk/projects/UNITE/, (10/03/2008).

Varian, H. (1992) Microeconomic Analysis, 3rd edition, Norton, New-York.

Verband Deutscher Verkehrsunternehmen (VDV-Statistik) (2006) VDV-Statistik 2006,

http://www.vdv.de/module/layout_upload/st2006_online.pdf, (10/03/2008).

Verhoef, E. T. (1995) The Demand Curve under Road Pricing and the Problem of Political

Feasibility: A Comment. Transportation Research Part A: Policy and Practice, Vol.

29(6), pp. 459-465.

173

Vickrey, W. S. (1948) Some Objections to Marginal-Cost Pricing. Journal of Political

Economy 56, pp. 218-238.

Vickrey, W. S. (1968) Congestion Charges and Welfare: Some Answers to Sharp's Doubts.

Journal of Transport Economics and Policy 2, pp. 107-118.

Viegas, J. M. (2001) Making urban road pricing acceptable and effective: Searching for

quality and equity in urban mobility. Transport Policy 8, pp. 289-294.

Walras, L. (1874) Eléments d’économie pure ou théorie de la richesse sociale. Corbaz,

Lausan. Elements of Pure Economics. Trans. (1954) by W. Jafle. London: Allen and

Unwin.

Walster, E., G. W. Walster, and E. Berscheid (1978) Equity theory and research. Boston:

Allyn and Bacon.

Walters, A. A., (1961) The Theory and Measurement of Private and Social Cost of

Highway Congestion. Econometrica, 29, pp. 676-99.

Webb, S. and B. Webb (1920) English local government: the story of the king’s highway.

174

175

8 Appendix

Appendix 1 List of core model equations

Source: Steininger and Friedl (2004).

Production

(1)

(

)

min , 1,...,35

j j j ij ij

X H A X a for j= =

(2)

( )

(

)

1 1

1 1,...,35

j j

j j j j

j j j j j

H L K for j

σ σ

σ σ σ σ

δ δ

−

− −

= + − =

(3)

(

)

min ,

pf pf pv pv

A A

T T T=

(4)

(

)

min

i i

X A

(5)

(

)

min ,

pv p kmp

i i

X A km A

(6)

(

)

min

i i

X A

Foreign Trade

(7)

(

)

1,...,35

o w

j j j j

EX EX P P for j

= =

(8)

(

)

1,...,35

o w

j j j j

M M P P for j

= =

Labour Market

(9)

≥ ⊥

low

w u

Household Demand

176

(10)

( )

(

)

,,, ,

, , , . .

-1

11 14

41 44

 

−

 

 

−

 

= 

 

 

 

 

= 

 

 

+ −

M O M

he r

Ce r

C C

he r e r h h

e r e r

e r e r e r e r e r

C C

h h h h h

e r

C T

h h

for h where e is the income category and r the resi

dential location attribute

h h

σ σ

δ δ

(11)

(

)

( )

, ,

, , ,

, ,

11 14

, ,

41 44

1 ,

 

−

 

 

−

 

 

= 

 

 

 

 

 

 

= =  

 

 

∑

M O M

X X

h h

e r e r

X X

h h

e r e r

e r e r e r

e r

c X c

h h i h i

h i e r

h h

with for h

h h

where e is the income category and r the resident

ial location attribute

σ σ

(12)

( )

(

)

, ,

-1 -1

, ,

, , , , ,

11 14

41 44

   

   

   

−

 

 

= + − +

 

 

 

 

= 

 

 

M O M

T T

h h

e r e r

T T T T

h h

e r h e r h

e r e r

e r e r e r e r e r

T p T u

h h h h h

e r

h h

for h

h h

where e is the income category and r the resident

ial location attribute

T T

σ σ σ σ

σ σ

δ δ

177

Appendix 2 Variables

Source: Steininger and Friedl (2004).

Factor demand

Production

Foreign trade

, M

Labour Market

low

Total labour demand

Total capital demand

Gross production of sector j

Capital input in sector j

Labour input in sector j

Factor aggregate in sector j

Leontief-input -output -coefficients in sector j

CES-distribution parameter in sector j

Elasticity of substitution in production between labour and

capital in sector j

Export of sector j

Import of sector j

Production price of goods aggregate X in sector j

World market price of goods aggregate M in sector j

Export and import quantities in sector j in the reference year

Foreign trade price elasticity of demand in sector j

Nominal wage rate

Lower bound on the real wage rate

Paasche index of the aggregate price level

Rate of unemployment

178

Transport

, A

pfi

, A

pvi

kmp

Consumption

h,iX

Private car passenger transport

Private car passenger transport production fixed input

Private car passenger transport production variable input

(directly kilometre dependent)

Public passenger transport

Leontief-input -output -coefficients in private car passenger

transport

Kilometre input coefficient in private car passenger transport

Leontief-input -output -coefficients in public transport

Vehicle kilometres driven in private car transport

Total Consumption of household type h

Consumption of non-transport goods of household h

Transport consumption of household h

CES-distribution parameter in consumption for household h

CES-distribution parameter in transport consumption for

household h

CES-distribution parameter in non-transport consumption for

household h

Elasticity of substitution between transport and non-transport

demand for household h

Elasticity of substitution between private car transport and

Household type can be distinguished by income level and/ or residential location attribute:

11 14

41 44

 

 

= 

 

 

M O M

e r

h h

h where e is the income category and r the residen

tial location attribute

h h

179

public transport demand for household h

Elasticity of substitution between non-transport goods in

household h consumption

180

Appendix 3 Gross domestic product calculation approaches for an input-output

table

Source: Eurostat (2008).

Production approach Income approach Expenditure approach

Total output at basic prices Compensations of employees Household final consumption expenditure

– Intermediate consumption + Other net taxes on production + NPISH final consumtion expenditure

– Taxes less subsides on products + Capital consumption + Government consumtion expenditure

+ Net operating surplus + Gross fixed capital formation

+ Changes in inventories

= Value added at basic prices = Value added at basic prices + Acquisitions less disposals of valuables

+ Exports of goods and services

+ Taxes less subsides on products + Taxes less subsides on products – Imports of goods and services

– Direct purchases abroad by residents

= Gross domestic products = Gross domestic products = Gross domestic products

181

Appendix 4 Input-output tables and data sources within the national and European

system of accounts

Source: Eurostat (2008).

(A)

European

System of

Accounts

ESA 1995

(B)

Eurostat

Manual of

Supply, Use &

Input-Output

Tables

(C)

System of

National

Accounts

SNA 1993

(D)

United

Nations

Handbook of

Input-Output

Tables

(20)

Consumer

Price

Statistics

European System of Accounts

(18)

Capital

Expenditure

Survey

Supply and

use tables at

purchasers'

prices

(unbalanced)

Valuation

matrices

(unbalanced)

(17)

Material

Input

Statistics

(13)

Census of

Agriculture

(12)

Agriculture

Statistics

(11)

Population

Census

(10)

Employment

Statistics

(9)

Monetary

Statistics

(8)

Financial

Statistics

(14)

Production

Statistics

Product by

product

input-output

tables

Industry by

industry

input-output

tables

Goods and

services

account

Production

account

Distribution

and use of

income

accounts

Accumulatio

accounts

(19)

Producer

Price

Statistics

Production

approach

GDP

(unbalanced)

Income

approach

GDP

(unbalanced)

Expenditure

approach

GDP

(unbalanced)

(15)

Foreign

Trade

Statistics

Supply and use tables at basic

prices Total

economy

Sector accounts

1. Non-financial corporations

2. Financial corporations

3. General government 4. Households

5. Non-profit institutions serving households

(16)

Balance of

Payments

Supply and

use tables at

purchasers'

prices

(balanced)

Valuation

matrices

(balanced)

(1)

Establishmen

Census

(3)

Annual

Business

Surveys

(4)

Construction

survey

(7)

Government

Expenditure

and

Revenue

(2)

Business

(6)

Consumer

Expenditure

Survey

(5)

Income

survey

Supply and use system Accounts

Balancing

Gross domestic product

(balanced)

182

Appendix 5 Household refund distribution in scenario B according to fuel tax

payments

Household fuel tax share in the overall fuel tax state contribution, in %

Income

HIQ1 6.9

HIQ2 23.5

HIQ3 29.8

HIQ4 39.7

Residential location

HRCent1 19.6

HRAgglo2 43.8

HRUrb3 22.9

HRRul4 13.7

Income and residential location

H1 1.9

H2 4.3

H3 5.6

H4 7.8

H5 1.3

H6 8.2

H7 13.8

H8 20.5

H9 2.2

H10 6.6

H11 6.6

H12 7.5

H13 1.5

H14 4.4

H15 3.8

H16 3.9

Sources: EVS 2003 (StaBuA, 2005), and own calculations.

183

Appendix 6 Household expenditures on fixed car use related components for

household categories

Transportation expenditure in million Euro

Residential category

Purchase of new and second-hand vehicles

Income quartlie

HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ1 730.6 763.7 724.5 544.4 2,763

HIQ2 1,988 3,771 2,862 1,968 10,590

HIQ3 3,026 6,914 2,222 1,488 13,650

HIQ4 4,372 10,611 5,061 2,158 22,203

Total 10,117 22,060 10,870 6,159 49,206

Spare parts and accessory

HIQ1 85.0 104.9 141.9 73.8 405.6

HIQ2 244.8 479.2 350.2 235.4 1,310

HIQ3 338.3 761.5 337.8 235.8 1,673

HIQ4 434.0 1,188 501.7 247.2 2,371

Total 1,102 2,534 1.332 792.3 5,760

Maintenance and repair

HIQ1 352.1 153.7 346.7 212.2 1,065

HIQ2 638.2 1,409 1,084 681.9 3,813

HIQ3 961.5 2,291 949.2 499.8 4,702

HIQ4 1,368 3,126 1,026 566.1 6,087

Total 3,320 6,980 3,406 1,960 15,666

Other services related to car use

HIQ1 173.0 109.7 105.3 59.3 447.3

HIQ2 170.5 356.9 242.1 136.4 905.9

HIQ3 228.2 493.0 206.1 118.1 1,045

HIQ4 332.3 717.6 234.2 127.6 1,412

Total 904.1 1,677.3 787.6 441.5 3,810

(Annual) Tax on motor vehicles

HIQ1 90.9 63.9 122.9 80.2 357.9

HIQ2 231.3 520.4 393.0 267.1 1,412

HIQ3 286.9 778.1 318.4 199.8 1,583

HIQ4 415.6 1,048 377.0 207.8 2,048

Total 1,025 2,410 1,211 754.8 5,401

Car insurance and related financial services

HIQ1 348.4 239.3 360.4 261.9 1,210

HIQ2 729.4 1,337 1,152 804.8 4,023

HIQ3 865.9 2,287 1,002 637.2 4,792

HIQ4 1,319 3,440 1,251 644.7 6,655

Total 3,262 7,303 3,766 2,349 16,680

Total fixed car travel expenditure

HIQ1 1,780 1,435 1,802 1,232 6,249

HIQ2 4,002 7,874 6,084 4,094 22,054

HIQ3 5,707 13,525 5,036 3,179 27,446

HIQ4 8,241 20,131 8,451 3,952 40,775

Total 19,730 42,965 21,372 12,456 96,524

Sources: EVS 2003 (StaBuA, 2005), and own calculations.

184

Appendix 7 Household expenditures on transportation as shares in the overall

consumption expenditure for household categories

Fix car travel related expenditures as total consumption expenditure shares in % for different household categories

Residential category

Total fix car travel expenditure

Income quartlie HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ1 3.5 2.8 5.1 5.4 3.9

HIQ2 8.6 8.8 9.8 10.5 9.3

HIQ3 10.1 10.1 10.1 11.3 10.2

HIQ4 11.4 11.6 14.8 14.3 12.3

Total 8.7 9.6 10.5 10.6 9.7

Total variable car travel expenditure for fuels

HIQ1 1.4 1.0 2.3 2.6 1.6

HIQ2 3.5 3.5 4.0 4.2 3.7

HIQ3 3.7 3.9 5.0 5.1 4.2

HIQ4 4.0 4.4 5.0 5.3 4.5

Total 3.2 3.7 4.2 4.4 3.8

Total car travel expenditure

HIQ1 4.8 3.7 7.5 8.0 5.5

HIQ2 12.0 12.3 13.8 14.7 13.0

HIQ3 13.8 14.0 15.1 16.4 14.4

HIQ4 15.4 16.1 19.8 19.7 16.9

Total 11.9 13.3 14.7 15 13.5

Total public transportation expenditure

HIQ1 2.2 2.0 1.3 1.1 1.8

HIQ2 1.3 0.8 0.6 0.5 0.8

HIQ3 1.1 0.7 0.6 0.6 0.7

HIQ4 1.2 0.7 0.5 0.5 0.8

Total 1.4 0.9 0.7 0.7 0.9

Total travel expenditure

HIQ1 7.1 5.7 8.8 9.1 7.3

HIQ2 13.4 13.1 14.4 15.3 13.9

HIQ3 14.9 14.7 15.7 17.0 15.2

HIQ4 16.6 16.8 20.2 20.2 17.6

Total 13.4 14.2 15.4 15.7 14.4

Sources: EVS 2003 (StaBuA, 2005), and own calculations.

185

Appendix 8 Household expenditures on fixed car use related components as shares

in household net income for household categories

Fix car travel related expenditures as net income shares in % for different household categories

Residential category

Purchase of new and second-hand vehicles

Income quartile HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ1 1.2 1.3 1.7 2.0 1.5

HIQ2 3.4 3.4 3.7 4.0 3.6

HIQ3 4.3 4.1 3.3 3.8 3.9

HIQ4 4.3 4.3 6.4 5.6 4.8

Total 3.5 3.8 4.1 4.0 3.8

Spare parts and accessory

HIQ1 0.1 0.2 0.3 0.3 0.2

HIQ2 0.4 0.4 0.5 0.5 0.4

HIQ3 0.5 0.4 0.5 0.6 0.5

HIQ4 0.4 0.5 0.6 0.6 0.5

Total 0.4 0.4 0.5 0.5 0.4

Maintenance and repair

HIQ1 0.6 0.3 0.8 0.8 0.6

HIQ2 1.1 1.3 1.4 1.4 1.3

HIQ3 1.4 1.3 1.4 1.3 1.3

HIQ4 1.3 1.3 1.3 1.5 1.3

Total 1.1 1.2 1.3 1.3 1.2

Other services related to car use

HIQ1 0.3 0.2 0.3 0.2 0.2

HIQ2 0.3 0.3 0.3 0.3 0.3

HIQ3 0.3 0.3 0.3 0.3 0.3

HIQ4 0.3 0.3 0.3 0.3 0.3

Total 0.3 0.3 0.3 0.3 0.3

(Annual) Tax on motor vehicles

HIQ1 0.1 0.1 0.3 0.3 0.2

HIQ2 0.4 0.5 0.5 0.5 0.5

HIQ3 0.4 0.5 0.5 0.5 0.5

HIQ4 0.4 0.4 0.5 0.5 0.4

Total 0.4 0.4 0.5 0.5 0.4

Car insurance and related financial services

HIQ1 0.6 0.4 0.9 1.0 0.6

HIQ2 1.3 1.2 1.5 1.7 1.4

HIQ3 1.2 1.3 1.5 1.6 1.4

HIQ4 1.3 1.4 1.6 1.7 1.4

Total 1.1 1.2 1.4 1.5 1.3

Total fixed car travel expenditure

HIQ1 2.9 2.4 4.3 4.6 3.3

HIQ2 6.9 7.2 7.9 8.4 7.5

HIQ3 8.0 7.9 7.4 8.2 7.9

HIQ4 8.1 8.2 10.4 10.3 8.8

Total 6.8 7.3 8.0 8.1 7.5

Sources: EVS 2003 (StaBuA, 2005), and own calculations.

186

Appendix 9 Household expenditures on fixed car use related components as shares

in household overall consumption expenditure for household categories

Fix car travel related expenditures as total consumption expenditure shares in % for different household categories

Residential category

Purchase of new and second-hand vehicles

Income quartile HRCent1 HRAgglo2 HRUrb3 HRRul4 Total

HIQ1 1.4 1.5 2.1 2.4 1.7

HIQ2 4.3 4.2 4.6 5.1 4.5

HIQ3 5.4 5.2 4.5 5.3 5.1

HIQ4 6.0 6.1 8.9 7.8 6.7

Total 4.5 4.9 5.3 5.2 4.9

Spare parts and accessory

HIQ1 0.2 0.2 0.4 0.3 0.3

HIQ2 0.5 0.5 0.6 0.6 0.6

HIQ3 0.6 0.6 0.7 0.8 0.6

HIQ4 0.6 0.7 0.9 0.9 0.7

Total 0.5 0.6 0.7 0.7 0.6

Maintenance and repair

HIQ1 0.7 0.3 1.0 0.9 0.7

HIQ2 1.4 1.6 1.7 1.8 1.6

HIQ3 1.7 1.7 1.9 1.8 1.8

HIQ4 1.9 1.8 1.8 2.1 1.8

Total 1.5 1.6 1.7 1.7 1.6

Other services related to car use

HIQ1 0.3 0.2 0.3 0.3 0.3

HIQ2 0.4 0.4 0.4 0.4 0.4

HIQ3 0.4 0.4 0.4 0.4 0.4

HIQ4 0.5 0.4 0.4 0.5 0.4

Total 0.4 0.4 0.4 0.4 0.4

(Annual) Tax on motor vehicles

HIQ1 0.2 0.1 0.4 0.4 0.2

HIQ2 0.5 0.6 0.6 0.7 0.6

HIQ3 0.5 0.6 0.6 0.7 0.6

HIQ4 0.6 0.6 0.7 0.8 0.6

Total 0.5 0.5 0.6 0.6 0.5

Car insurance and related financial services

HIQ1 0.7 0.5 1.0 1.2 0.8

HIQ2 1.6 1.5 1.9 2.1 1.7

HIQ3 1.5 1.7 2.0 2.3 1.8

HIQ4 1.8 2.0 2.2 2.3 2.0

Total 1.4 1.6 1.8 2.0 1.7

Total fixed car travel expenditure

HIQ1 3.5 2.8 5.1 5.4 3.9

HIQ2 8.6 8.8 9.8 10.5 9.3

HIQ3 10.1 10.1 10.1 11.3 10.2

HIQ4 11.4 11.6 14.8 14.3 12.3

Total 8.7 9.6 10.5 10.6 9.7

Sources: EVS 2003 (StaBuA, 2005), and own calculations.

187

Appendix 10 Accessibility of centres and functional urban areas in Germany by car,

German Federal Office for Building and Regional Planning (BBSR)

The spatial distribution of a car accessibility index calculated by the BBSR as shown

(below) accounts more for the proximity in terms of travel time than travel distance. The

structural picture reveals the rather heterogeneous, mono- and polycentric distribution of

functional urban areas in Germany and the functional linkages implied by the network of

national road infrastructure. In Germany road infrastructure has in general a significant

influence on accessibility. 58 % of the German country area can be attributed to rural

regions or peripheries. Their population densities are about 100 inhabitants per km

compared to the national average of about 230 people per km

. Nevertheless, these areas

comprise one quarter of the entire German population. The resulting spatial dispersion of

scarcely populated rural areas in Germany contributes to the car dependency observed in

some places (Schuert et al., 2005).

188

189

Appendix 11 GAMS code of the German Road Travel Policy Model (GRTPM), 2002

* Multihousehold hh version GRTPM

* Year of calculation: 2002

* Dissertation Kalinowska Dominika

scalar urbase unemployment rate base year /9.41/

uabase unemployed in 1000 base year /4061/

lbase tot labor force avail incl unempl in 1000

/43157/

emp total labor force employed base year in 1000

/39096/;

* PUBLIC FINANCE

scalar lwtaxr labor wage tax rate /0.930681064/

ktaxr capital revenue tax rate /0.289784016 /

pubdef public deficit in mio euro /-64300/;

2222222222222222222222222222222222222222222222222222

$ontext

LAWI

KOHLE

OELBB

OELVER

ELEK

WASSER

EISEN

STEIN

CHEMIE

METALL

MASCH

BUEROM

ELEINR

FAHRZ

NAHR

TEXTIL

HOLZ

PAPIER

VERLAG

GUMMI

RECYC

SPROD

BAU

HANDEL

GAST

VERK

SUL

SVERK

KOMM

GELD

190

REAL

DATEN

FUE

SODIEN

NMDIEN

UALLO

$offtext

22222222222222222222222222222222222222222222

* D E F I N I T I O N S

22222222222222222222222222222222222222222222

set esps economic sectors and primary factors

/LAWI, KOHLE, OELBB, OELVER, ELEK, WASSER, EISEN, STEIN, CHEMIE, METALL,

MASCH, BUEROM, ELEINR, FAHRZ, NAHR, TEXTIL, HOLZ, PAPIER, VERLAG,

GUMMI, RECYC, SPROD, BAU, HANDEL, GAST, VERK, SUL, SVERK, KOMM,

GELD, REAL, DATEN, FUE, SODIEN, NMDIEN, ITAX, LTAX, KTAX,

L, K, TXTRNS, UBEN, W, FX, ITAXRT, OUTPUT, EX, IM/,

espw economic sectors and agents

/LAWI, KOHLE, OELBB, OELVER, ELEK, WASSER, EISEN, STEIN, CHEMIE, METALL,

MASCH, BUEROM, ELEINR, FAHRZ, NAHR, TEXTIL, HOLZ, PAPIER, VERLAG,

GUMMI, RECYC, SPROD, BAU, HANDEL, GAST, VERK, SUL, SVERK, KOMM,

GELD, REAL, DATEN, FUE, SODIEN, NMDIEN, W, CONS, GOV, EX, IM,

OUTPUT, L, K, ITAXRT, BESCH/,

es(esps) economic sectors

/LAWI, KOHLE, OELBB, OELVER, ELEK, WASSER, EISEN, STEIN, CHEMIE, METALL,

MASCH, BUEROM, ELEINR, FAHRZ, NAHR, TEXTIL, HOLZ, PAPIER, VERLAG,

GUMMI, RECYC, SPROD, BAU, HANDEL, GAST, VERK, SUL, SVERK, KOMM,

GELD, REAL, DATEN, FUE, SODIEN, NMDIEN/,

ess(espw) economic sectors again

/LAWI, KOHLE, OELBB, OELVER, ELEK, WASSER, EISEN, STEIN, CHEMIE, METALL,

MASCH, BUEROM, ELEINR, FAHRZ, NAHR, TEXTIL, HOLZ, PAPIER, VERLAG,

GUMMI, RECYC, SPROD, BAU, HANDEL, GAST, VERK, SUL, SVERK, KOMM,

GELD, REAL, DATEN, FUE, SODIEN, NMDIEN/,

hh household types

/h1, h2, h3, h4, h5, h6, h7, h8, h9, h10, h11, h12, h13, h14, h15, h16 /,

*households

hs household sums

/all_hh,he1,he2,he3,he4,hr1,hr2,hr3,hr4/,

hzu(hs,hh) / all_hh.(h1, h2, h3, h4, h5, h6, h7, h8, h9, h10, h11, h12,

h13, h14, h15, h16)

he1.(h1, h5, h9, h13)

he2.(h2, h6, h10, h14)

he3.(h3, h7, h11, h15)

he4.(h4, h8, h12, h16)

hr1.(h1, h2, h3, h4)

hr2.(h5, h6, h7, h8)

hr3.(h9, h10,h11, h12)

hr4.(h13,h14,h15, h16)/,

191

he1(hh) household income type1

/h1, h5, h9, h13/,

he2(hh) household income type2

/h2, h6, h10, h14/,

he3(hh) household income type3

/h3, h7, h11, h15/,

he4(hh) household income type4

/h4, h8, h12, h16/,

hr1(hh) household regional type1

/h1, h2, h3, h4/,

hr2(hh) household regional type2

/h5, h6, h7, h8/,

hr3(hh) household regional type3

/h9, h10, h11, h12/,

hr4(hh) household regional type4

/h13, h14, h15, h16/,

he household income types

/he1, he2, he3, he4/,

hr household regional types

/hr1, hr2, hr3, hr4/;

ALIAS (ES,SS);

2222222222222222222222222222222222222222222222222222222222222222222222

* E C O N O M I C D A T A

2222222222222222222222222222222222222222222222222222222222222222222222

table sam(esps,espw) benchmark social accounting matrix

$ondelim

$include d:\diss\ARPM Model\samd2002.csv

$offdelim

display sam;

set r(esps)

/LAWI, KOHLE, OELBB, OELVER, ELEK, WASSER, EISEN, STEIN, CHEMIE,

METALL,

MASCH, BUEROM, ELEINR, FAHRZ, NAHR, TEXTIL, HOLZ, PAPIER, VERLAG,

GUMMI, RECYC, SPROD, BAU, HANDEL, GAST, VERK, SUL, SVERK, KOMM,

GELD, REAL, DATEN, FUE, SODIEN, NMDIEN, ITAX, LTAX, KTAX,

L, K, TXTRNS, UBEN, W, FX, OUTPUT/,c(espw)

/LAWI, KOHLE, OELBB, OELVER, ELEK, WASSER, EISEN, STEIN, CHEMIE,

METALL,

MASCH, BUEROM, ELEINR, FAHRZ, NAHR, TEXTIL, HOLZ, PAPIER, VERLAG,

GUMMI, RECYC, SPROD, BAU, HANDEL, GAST, VERK, SUL, SVERK, KOMM,

GELD, REAL, DATEN, FUE, SODIEN, NMDIEN, W, CONS, GOV, EX, IM,

OUTPUT/;

parameter rchk(r) row sum check

cchk(c) column sum check;

rchk(r) = sum(c, sam(r,c));

cchk(c) = sum(r, sam(r,c));

table sams(es,es) input table per sector

LOOP(es,LOOP(ss, sams(ss,es)=(sum(ess,sam(ss,ess)$(ord(ess)=ord(es))))))

192

22222222222222222222222222222222222222222222222222222222222222222222222

* MULTIHOUSEHOLD INCOME DISTRIBUTION

22222222222222222222222222222222222222222222222222222222222222222222222

*16 new hh categories as follows

*In Kernstaedten< 500 TEW/u. 1500 h1

*In Kernstaedten< 500 TEW/1500-u.2600 h2

*In Kernstaedten< 500 TEW/2600-u.3600 h3

*In Kernstaedten< 500 TEW/3600 h4

*Agglom oh.Kernstaedte>500TEW/u.1500 h5

*Agglom oh.Kernstaedte>500TEW/1500-u.2600 h6

*Agglom oh.Kernstaedte>500TEW/2600-u.3600 h7

*Agglom oh.Kernstaedte>500TEW/3600+ h8

*verstaedt.Raeume/u.1500 h9

*verstaedt.Raeume/1500-u.2600 h10

*verstaedt.Raeume/2600-u.3600 h11

*verstaedt.Raeume/3600+ h12

*laendl.Raeume/u.1500 h13

*laendl.Raeume/1500-u.2600 h14

*laendl.Raeume/2600-u.3600 h15

*laendl.Raeume/3600+ h16

*total income incorporates selfemployed and employees

parameter incs(hh) share of total income per hh

/h1 0.046775117

h2 0.044767507

h3 0.054814582

h4 0.078308642

h5 0.046817928

h6 0.08485037

h7 0.131373693

h8 0.188990716

h9 0.032021855

h10 0.059536605

h11 0.05261998

h12 0.060985583

h13 0.020858652

h14 0.037630954

h15 0.030092354

h16 0.029557284 /;

parameter incls(hh) share of labour income per hh

/h1 0.014128817

h2 0.042987016

h3 0.04476612

h4 0.112594548

h5 0.019543136

h6 0.062671388

h7 0.116643931

h8 0.266955381

h9 0.013081157

h10 0.039955945

h11 0.066044755

193

h12 0.085483168

h13 0.008848666

h14 0.026795984

h15 0.037595981

h16 0.041904007 /;

parameter ubens(hh) share of unemployed transfer income per hh

/h1 0.086020576

h2 0.078657124

h3 0.055243187

h4 0.036505936

h5 0.08875567

h6 0.107662339

h7 0.087543123

h8 0.108706424

h9 0.071250367

h10 0.075176916

h11 0.035015962

h12 0.030805725

h13 0.053374574

h14 0.048256089

h15 0.019780496

h16 0.01724549 /;

parameter transfs(hh) share of transfers income per hh

/h1 0.146448135

h2 0.039522829

h3 0.082206489

h4 0.019764343

h5 0.109499189

h6 0.105249511

h7 0.158270127

h8 0.043497748

h9 0.085661845

h10 0.079260119

h11 0.01720345

h12 0.009928888

h13 0.050918837

h14 0.043375241

h15 0.005940016

h16 0.003253232 /;

*NEU090709

parameter ks(hh) share of capital income per hh

/h1 0.040395493

h2 0.045704825

h3 0.062946637

h4 0.07809991

h5 0.036546907

h6 0.090294066

h7 0.150426334

h8 0.18457952

h9 0.023643368

h10 0.06225439

194

h11 0.048943646

h12 0.061348323

h13 0.01701506

h14 0.040860167

h15 0.027117022

h16 0.029824332 /;

parameter fxs(hh) share of foreign exchange income per hh

/h1 0.161407197

h2 0.037914051

h3 0.125314168

h4 0.054621517

h5 0.063942647

h6 0.07495688

h7 0.216406296

h8 0.086412885

h9 0.069888877

h10 0.044075831

h11 -0.009550738

h12 0.020401341

h13 0.043387442

h14 0.023455465

h15 -0.021518899

h16 0.00888504 /;

22222222222222222222222222222222222222222222222222222222222222222222222

* SAM ADJUSTMENT

22222222222222222222222222222222222222222222222222222222222222222222222

parameter kfzsteuer(hh) car tax per hh

/h1 90.868

h2 231.321

h3 286.896

h4 415.613

h5 63.891

h6 520.411

h7 778.117

h8 1047.807

h9 122.912

h10 393.012

h11 318.424

h12 376.950

h13 80.185

h14 267.095

h15 199.760

h16 207.769 /;

parameter moest(hh) energy tax per hh

/h1 444.915

h2 1023.445

h3 1325.382

h4 1837.088

195

h5 315.961

h6 1936.243

h7 3260.101

h8 4856.374

h9 514.848

h10 1565.248

h11 1569.780

h12 1779.233

h13 365.662

h14 1035.338

h15 907.527

h16 926.704 /;

parameter oevexp(hh) variable public transport expenditure per hh

/h1 1470.036

h2 777.158

h3 784.409

h4 1083.058

h5 1321.336

h6 925.970

h7 1175.035

h8 1628.086

h9 602.604

h10 481.844

h11 349.595

h12 344.734

h13 324.140

h14 270.303

h15 210.141

h16 189.975 /;

parameter inpfz(hh) production input from vehicle prduction

sector per hh

/h1 541.904

h2 1474.276

h3 2244.736

h4 3243.163

h5 566.425

h6 2797.244

h7 5128.568

h8 7870.527

h9 537.414

h10 2123.023

h11 1647.955

h12 3753.873

h13 403.781

h14 1460.038

h15 1103.675

h16 1600.764 /;

parameter inpha(hh) production input from trade etc. per hh

/h1 63.018

h2 181.590

h3 250.953

196

h4 321.932

h5 77.835

h6 355.422

h7 564.827

h8 881.188

h9 105.269

h10 259.781

h11 250.534

h12 372.096

h13 54.749

h14 174.633

h15 174.918

h16 183.388 /;

parameter inpge(hh) production input from finance and insurance per hh

/h1 348.385

h2 729.367

h3 865.853

h4 1318.850

h5 239.289

h6 1336.873

h7 2286.981

h8 3440.181

h9 360.428

h10 1152.320

h11 1002.348

h12 1250.917

h13 261.894

h14 804.822

h15 637.204

h16 644.747 /;

parameter inpso(hh) production input from aux. services per hh

/h1 161.144

h2 158.829

h3 212.540

h4 309.497

h5 102.196

h6 332.401

h7 459.174

h8 668.389

h9 98.034

h10 225.449

h11 191.913

h12 218.169

h13 55.201

h14 127.080

h15 110.027

h16 118.881 /;

parameter inpoe(hh) production input from fuel per hh

/h1 261.300

h2 601.071

h3 778.399

197

h4 1078.925

h5 185.564

h6 1137.158

h7 1914.663

h8 2852.156

h9 302.371

h10 919.272

h11 921.935

h12 1044.946

h13 214.754

h14 608.055

h15 532.992

h16 544.255 /;

parameter inphav(hh) production inpurt from trade to variable

cost per hh

/h1 261.149

h2 473.352

h3 713.208

h4 1014.772

h5 114.021

h6 1045.238

h7 1699.309

h8 2318.786

h9 257.185

h10 804.197

h11 704.024

h12 761.166

h13 157.408

h14 505.767

h15 370.747

h16 419.914 /;

parameter oeltaxr(hh) energy tax rate on fuel use per hh;

oeltaxr(hh)=(moest(hh)+inpoe(hh))/inpoe(hh)-1;

* sum fixed and variable car use expenditure and variable public

transport expenditure

parameter vausg(hh) total private hh transport expenditure per

mivfix(hh) fixed car expenditure per hh

mivvar(hh) variable car expenditure per hh

mivcar(hh) total car expenditure per hh

mivfix(hh)=kfzsteuer(hh)+inpfz(hh)+inpha(hh)+inpge(hh)+inpso(hh);

mivvar(hh)=moest(hh)+inpoe(hh)+inphav(hh);

vausg(hh)=oevexp(hh)+mivfix(hh)+mivvar(hh);

mivcar(hh)=mivfix(hh)+mivvar(hh);

parameter capo(hh) calibrated price for oelverbr;

capo(hh)=1+oeltaxr(hh);

22222222222222222222222222222222222222222222222222222222222222222222222

* ROAD PRICING BASE DATA

22222222222222222222222222222222222222222222222222222222222222222222222

198

sam("OELVER","W") =sam("OELVER","W")- (sum(hh, inpoe(hh)));

sam("FAHRZ","W") =sam("FAHRZ","W")- (sum(hh, inpfz(hh)));

sam("HANDEL","W") =sam("HANDEL","W")- (sum(hh, inpha(hh)+inphav(hh)));

sam("VERK","W") =sam("VERK","W")- (sum(hh, oevexp(hh)));

sam("GELD","W") =sam("GELD","W")- (sum(hh, inpge(hh)));

sam("SODIEN","W") =sam("SODIEN","W")- (sum(hh, inpso(hh)));

* REFUND OF ROAD PRICING REVENUE

scalar netrev road pricing revenue net of system costs;

netrev=0.85;

scalar hhref share of road pricing revenues refunded to hh;

hhref=1/3;

* hhref=0.00001;

scalar revshare share of road pricing revenues used for oev;

revshare=0.5;

parameter ivlevel level of car transport relative to base;

ivlevel=1.0;

parameter rotax road pricing tax level in Euro per km (mio Euro

per mio km);

rotax=0.000000001;

parameter oetax hypothetic oev tax level in Euro;

oetax=0.000000001;

parameter mivkm(hh) car km per hh in mio

/h1 10721.794

h2 22023.285

h3 26403.228

h4 44656.949

h5 7988.711

h6 32541.210

h7 68914.639

h8 104213.152

h9 10450.981

h10 31939.475

h11 31527.906

h12 37304.760

h13 11952.024

h14 18684.331

h15 16858.702

h16 16601.439 /;

parameter oevkm(hh) public travel pkm per hh in mio

/h1 12493.715

h2 7945.830

h3 13480.591

h4 12267.291

h5 9272.907

h6 8131.870

h7 15029.183

h8 17209.019

h9 6256.236

199

h10 8125.774

h11 5468.459

h12 3848.100

h13 2427.785

h14 5247.550

h15 2771.863

h16 3167.449 /;

parameter re1(hh) refund share per hh

/ h1 0.0625

h2 0.0625

h3 0.0625

h4 0.0625

h5 0.0625

h6 0.0625

h7 0.0625

h8 0.0625

h9 0.0625

h10 0.0625

h11 0.0625

h12 0.0625

h13 0.0625

h14 0.0625

h15 0.0625

h16 0.0625 /;

*different revenue redistribution scenario

*parameter re1(hh) refund share per hh

* / h1 0.019

* h2 0.043

* h3 0.056

* h4 0.087

* h5 0.013

* h6 0.082

* h7 0.138

* h8 0.205

* h9 0.022

* h10 0.066

* h11 0.066

* h12 0.075

* h13 0.015

* h14 0.044

* h15 0.038

* h16 0.039 /;

parameter CO

emifctr(hh) CO

emission factors

/ h1 0.213

h2 0.215

h3 0.215

h4 0.22

h5 0.211

h6 0.22

h7 0.218

h8 0.216

200

h9 0.213

h10 0.217

h11 0.216

h12 0.214

h13 0.21

h14 0.216

h15 0.214

h16 0.223 /;

scalar ckmmax car km highest possible value /1000000000/;

parameter ckmbase car km base year 2002;

ckmbase=sum(hh,mivkm(hh));

scalar oekmmax public travel pkm highest possible value /1000000000/;

parameter oekmbase public travel pkm base year 2002;

oekmbase=sum(hh,oevkm(hh));

22222222222222222222222222222222222222222222222222222222222222222222222

* F O R E I G N T R A D E P A R A M E T E R S

22222222222222222222222222222222222222222222222222222222222222222222222

parameter telas(es) Elast of substitution Armington from Reinert and

Holst

/LAWI 1.9

KOHLE 1.5

OELBB 1.5

OELVER 1.5

ELEK 1.5

WASSER 1.5

EISEN 1.7

STEIN 1.7

CHEMIE 0.7

METALL 0.8

MASCH 0.85

BUEROM 0.85

ELEINR 1.7

FAHRZ 1.4

NAHR 1.5

TEXTIL 1.17

HOLZ 1.6

PAPIER 1.4

VERLAG 1.4

GUMMI 1.55

RECYC 1.6

SPROD 1.6

BAU 0.5

HANDEL 1.0

GAST 1.0

VERK 1.9

SUL 1.9

SVERK 1.5

KOMM 1.0

GELD 0.5

REAL 0.5

201

DATEN 0.5

FUE 0.5

SODIEN 0.8

NMDIEN 0.4/;

22222222222222222222222222222222222222222222222222222222222222222222222

* L A B O R M A R K E T

22222222222222222222222222222222222222222222222222222222222222222222222

parameter urate unemployment rate

refurate reference is base case unemployment rate

uabs unemployed in 1000

unemp unemployment base year in foregone wage

earnings

emplchange change in employment absolute

usubsr unemployment subsidy rate

pusb price incl unemployment subsidy rate;

unemp=((-sam("L","Cons"))/emp)*uabase;

usubsr=((-sam("UBEN","CONS"))/unemp)*(-1);

pusb=1+usubsr;

222222222222222222222222222222222222222222222222222222222222222222222222222222

* G O V E R N M E N T R E V E N U E A N D E X P E N D I T U R E

222222222222222222222222222222222222222222222222222222222222222222222222222222

scalar nettransf net direct tax and transf payments plus public

saving

ubpwork unemployment benefit per worker

govdem government demand;

parameter pcalibl price level for calibration of factor labor

pcalibk price level for calibration of factor capital;

pcalibl=1+lwtaxr;

pcalibk=1+ktaxr;

govdem=sum(es,sam(es,"GOV"));

nettransf=(-sam("TXTRNS","CONS"));

ubpwork=(-sam("UBEN","CONS"))/uabase;

parameter def public deficit

defref public deficit reference base year level;

def=pubdef;

defref=def;

parameter tix(es) indirect tax rate net of subsidy;

tix(es)=sam(es,"ITAXRT");

22222222222222222222222222222222222222222222222222222222222222222222222

* F U R T H E R V A R I A B L E D E F I N I T I O N

22222222222222222222222222222222222222222222222222222222222222222222222

parameter opchp(es) sectoral output change in percent

test capital supply factor

dschp(es) sectoral output direct change first round in

percent

chemp(es) absolute change in sectoral employment

empchcorr correction term change in employment rounding

error

202

chex(es) change in exports

chim(es) change in imports

chva(es) change in value added per sector

tdirch direct tax revenue change absolute

tindch indirect tax revenue change absolute

ubench change in public unemployment benefit

subsidies absol

govch change in governement expenditure absolute

fch change in capital account

pxch(es) change in domestic production prices in %

pgch(es) change in domestic sales prices in %

pkch change in capital price

dirch(es) direct change first round in sectoral demand

dirchp(es) direct change first round in sectoral demand

in %

emivkm car km total

eoevkm public travel km total

mivch relative change in car travel km in %

oevch relative change in oev km in %

consch(hh) change in household consumption transport

vch(hh) change in total transport expenditure

exoev(hh) expenditure for public travel per hh group

exoevch(hh) change in expenditure for public travel per

exiv(hh) expenditure for travel per household group

exivch(hh) expenditure for travel change in %

eivkm(hh) total km car travel per household group

eovkm(hh) total km public travel per household group

hmivch(hh) relative change in miv km in % per household

group

hoevch(hh) relative change in oev km in % per household

group

rprev road pricing revenue total;

22222222222222222222222222222222222222222222222222222222222222222222222

* M P S / G E M O D E L

22222222222222222222222222222222222222222222222222222222222222222222222

$ontext

$model:rp_G16

$sectors:

x(es) ! output of domestic production sectors

y(es) ! level of domestic supply of domestic production

m(es) ! level of sectoral imports

ex(es) ! level of sectoral exports

g(es) ! level of domestic composite good supply

w ! Hicksian welfare index

v(hh) ! private transport demand bundle

iv(hh) ! car transport

ivv(hh) ! car transport variable components

ivf(hh) ! car transport fixed components

oev(hh) ! public transport

$commodities:

203

px(es) ! market price for sectoral domestic production

py(es) ! market price of domestic supply of domestic

production

pex(es) ! export price

pm(es) ! import price

pg(es) ! domestic composite good price

pl ! wage rate

pk ! capital rent

pw ! price index for welfare without transp

(expenditure function)

pv(hh) ! price index of passenger transport

piv(hh) ! price index of car tranpsort

pivv(hh) ! price index car transport variable components

pivf(hh) ! price index car transport fixed components

poev(hh) ! price index for public transport

pfx ! price foreign exchange (real terms of trade)

ptax ! price of tax payments

plinc ! price index of labour income

protax ! car road pricing tax level

poetax ! hypothetic publi ctransport pricing tax level

pkfzst ! car tax price index

$consumers:

cons(hh) ! representative agent

lmarket ! labour market overall

govt ! government

schick ! use of rp revenues agency

rps ! use of rp system expenditures

$auxiliary:

u ! unemployment rate

ufkmro ! restriction on car km

upkmoe ! restriction on public travel pkm

$prod:x(es) s:0 elk(s):0

o:px(es) q:(-sam(es,"OUTPUT")) A:govt T:TIX(es)

i:pg(ss) q:sams(ss,es)

i:pl q:sam(es,"L") P:pcalibl A:govt T:LWTAXR

elk:

i:pk q:sam(es,"K") P:pcalibk A:govt T:KTAXR

elk:

$prod:y(es) t:telas(es)

o:py(es) q:((-sam(es,"OUTPUT"))-sam(es,"EX"))

o:pex(es) q:sam(es,"EX")

i:px(es) q:(-sam(es,"OUTPUT"))

$prod:g(es) s:telas(es)

o:pg(es) q:((-sam(es,"OUTPUT"))-sam(es,"EX")-

sam(es,"IM"))

i:py(es) q:((-sam(es,"OUTPUT"))-sam(es,"EX"))

i:pm(es) q:(-sam(es,"IM"))

$prod:m(es)

204

o:pm(es) q:(-sam(es,"IM"))

i:pfx q:(-sam(es,"IM"))

$prod:ex(es)

o:pfx q:sam(es,"EX")

i:pex(es) q:sam(es,"EX")

$prod:iv(hh) s:0

o:piv(hh) q:(mivfix(hh)+mivvar(hh))

i:pivf(hh) q:mivfix(hh)

i:pivv(hh) q:mivvar(hh)

$prod:ivf(hh) s:0

o:pivf(hh) q:mivfix(hh)

i:pg("FAHRZ") q:inpfz(hh)

i:pg("HANDEL") q:inpha(hh)

i:pg("GELD") q:inpge(hh)

i:pg("SODIEN") q:inpso(hh)

i:pkfzst q:kfzsteuer(hh)

$prod:ivv(hh) s:0

o:pivv(hh) q:mivvar(hh)

i:pg("OELVER") q:inpoe(hh) P:capo(hh) A:govt T:oeltaxr(hh)

i:pg("HANDEL") q:inphav(hh)

i:protax q:mivkm(hh)

$prod:oev(hh) s:0

o:poev(hh) q:oevexp(hh)

i:pg("VERK") q:oevexp(hh)

i:poetax q:oekmbase

*Welfare

$prod:w s:1

o:pw q:(-sam("W","W")-sum(hh,vausg(hh)-

kfzsteuer(hh)-moest(hh)))

i:pg(es) q:(sam(es,"W"))

$prod:v(hh) s:0.635

o:pv(hh) q:vausg(hh)

i:piv(hh) q:(mivfix(hh)+mivvar(hh))

i:poev(hh) q:oevexp(hh)

$demand:cons(hh) s:0.275

d:pw q:((sam("W","CONS")*incs(hh))-

vausg(hh)+kfzsteuer(hh)+moest(hh))

d:pv(hh) q:vausg(hh)

e:plinc q:((-sam("L","CONS"))*incls(hh))

e:pk q:((-sam("K","CONS"))*ks(hh))

e:pfx q:((-sam("FX","CONS"))*fxs(hh))

e:ptax

q:(nettransf*transfs(hh)+kfzsteuer(hh)+moest(hh))

e:ptax q:((-sam("UBEN","CONS"))*ubens(hh)) R:u

e:protax q:(re1(hh)*hhref*netrev*ckmmax)

205

e:protax q:(re1(hh)*hhref*netrev*(ckmbase-ckmmax))

R:ufkmro

$demand:lmarket

d:plinc q:(-sam("L","CONS"))

e:pl q:((-sam("L","CONS"))+unemp)

e:pl q:(-unemp) R:u

$demand:govt

d:pg(es) q:sam(es,"GOV")

e:ptax q:(-nettransf-sum(hh,moest(hh)+kfzsteuer(hh)))

e:ptax q:sam("UBEN","CONS") R:u

e:pkfzst q:(sum(hh,kfzsteuer(hh)))

*no system costc included, revshare for public transport sector

$demand:schick

d:pg("BAU") q:((1-revshare)*0.9)

d:pg("NMDIEN") q:((1-revshare)*0.1)

d:pg("VERK") q:(revshare*0.9)

d:pg("NMDIEN") q:(revshare*0.1)

e:protax q:((1-hhref)*netrev*ckmmax)

e:protax q:((1-hhref)*netrev*(ckmbase-ckmmax))

R:ufkmro

e:poetax q:oekmmax

e:poetax q:(oekmbase-oekmmax)

R:upkmoe

*redistribution of system costs

$demand:rps

d:pg("GELD") q:(1/3)

d:pg("ELEINR") q:(1/3)

d:pl q:(1/3)

e:protax q:((1-netrev)*ckmmax)

e:protax q:((1-netrev)*(ckmbase-ckmmax))

R:ufkmro

$constraint:u

pl=E=pw;

$constraint:ufkmro

protax=G=rotax;

$constraint:upkmoe

poetax=G=oetax;

$REPORT:

V:domestic_output(es) O:px(es) PROD:x(es)

V:domestic_supply(es) O:py(es) PROD:y(es)

V:domestic_composite_supply(es) O:pg(es) PROD:g(es)

V:import(es) O:pm(es) PROD:m(es)

V:export(es) O:pfx PROD:ex(es)

V:car_transport(hh) O:piv(hh) PROD:iv(hh)

V:car_transport_fix(hh) O:pivf(hh) PROD:ivf(hh)

V:car_transport_var(hh) O:pivv(hh) PROD:ivv(hh)

206

V:public_transport(hh) O:poev(hh) PROD:oev(hh)

V:welfare_cons O:pw PROD:w

V:welfare_trans(hh) O:pv(hh) PROD:v(hh)

V:wcons(hh) W:cons(hh)

V:revenues_rp W:schick

V:GovDemand(es) D:pg(es) DEMAND:govt

$offtext

$sysinclude mpsgeset rp_G16

plinc.l=1;

ptax.fx=1;

*foreign price as numeraire

pfx.fx=1;

pkfzst.fx=1;

222222222222222222222222222222222222222222222222222222222222222222222222

* B A S E R U N

222222222222222222222222222222222222222222222222222222222222222222222222

pex.l(es)=1.0;

pm.l(es)=1.0;

u.l=1.0;

rotax=0.0000000001;

protax.l=0.0000000001;

oetax=0.0000000001;

poetax.l=0.0000000001;

ufkmro.l=1;

upkmoe.l=1;

rp_G16.iterlim=9999;

OPTION MCP = PATHC;

$include rp_G16.GEN

solve rp_G16 using mcp;

* O U T P U T

parameter gdpfactor gdp factor definition

govprice price index public demand

gdpuse gdp use definition

gdpref gdp adjusted for welfare price index

gdpbase gdp base level

govpbase govprice base

wcoref(hh) welfare index consumers hh reference level of

base year

gdpgrowth growth rate of gdp in use definition

base(hh) C02 base run

basecar(hh) CO

car base run

basepub(hh) C02pub base run

mivkmges overall travel km

mivkmbase(hh) car km base

oevkmbase(hh) public travel km base

vbase travele exp base

vausgg(hh) travel exp

207

oevexpbase(hh) public transport exp base

mivfixbase(hh) fixed car exp base

mivvarbase(hh) Varcarausgaben base

mivcarbase (hh) total car exp base

hhilabor(hh) hh income labour

hhicapit(hh) hh income capital

hhitrans(hh) hh income transfers

hhitransbase(hh) hh income transfers base

consbase(hh) consbase

(hh) CO

passenger travel

car(hh) C02 car

pub(hh) C02 pub ;

2222222222222222222222222222222222222222222222222222222222222222222222

* C O U N T E R F A C T U A L

2222222222222222222222222222222222222222222222222222222222222222222222

rotax=0.05;

protax.l=0.05;

rp_G16.iterlim=9999

$include rp_G16.GEN

solve rp_G16 using mcp;

208