Determinants and effects of infomobility at the city level

2nd International EIBURS-TAIPS conference on:“Innovation in the public sector

and the development of e-services”

University of UrbinoApril 18th-19th, 2013

Determinants and effects of infomobility at the city level

Davide Arduini, Marco Biagetti, Luigi Reggi and Paolo SeriEIBURS-TAIPS team, University of Urbino

2nd International EIBURS-TAIPS conference on:

““““Innovation in the public sector

and the development of e-services””””

Determinants and effects of infomobility

at the city-level

1

Davide Arduini, Marco Biagetti, Luigi Reggi, Paolo SeriEIBURS-TAIPS team

[email protected]

University of Urbino

April 19th, 2013

Plan of the talk

Research questions: 1) developing a model to explore theinfluence of some urban characteristics on theprovision/diffusion of Infomobility services; 2) analysing therelationships between urban pollution and ITS development

• Definition of Infomobility/Intelligent Transport Systems(ITS)

2

(ITS)

• Literature review

• Data and the econometric model

• Results

• Conclusion

Definition

• The concept of Infomobility/Intelligent Transport Systems (ITS), provided by theEuropean Commission (2003): "Intelligent Transport Systems: Intelligence at the

Service of Transport Networks“, include the following systems:

1) Advanced information for users; 2) Traffic control, navigation surveillance and guidance; 3)

Accident management; 4) Vehicle safety and control systems, as much as electronic payment andenforcement; 5) Operation of green zones/low emission zones; 6) Intermodality for bothpassenger and freight transport; 7) Interoperability standards, e.g. for ticketing

3

passenger and freight transport; 7) Interoperability standards, e.g. for ticketing

• The availability and adoption of ITS/Infomobility applications not only provides new

and flexible transport services but also a range of information services that have the

potential to increase the accessibility and usability of transport services, reduce

inequalities and increase economic participation and access to public services

� The combination of transportation accessibility, usability and availability culminatein the increased capacity of all citizens to participate in the local economy, accesspublic services and to be active members in their community

� ICT and Intelligent Transport Systems are improving all of these areas and arebreaking down geographical barriers as well

Literature review of Smart Cities (1/5)

• An increasing literature (Caragliu, Del Bo, Nijkamp, 2011; Arribas, Kourtit, Nijkamp,

2012; Deakin, 2012; Lombardi, Giordano, Farouh, Yousef, 2012) has highlighted that

there are several urban characteristics which are described in relation to the concept

“Smart City”: a) smart economy (related to competitiveness); b) smart people (related

to human capital); c) smart governance (related to participation); d) smart

environment (related to natural resources); e) smart living (related to the quality of

life)

• “Smart City” is furthermore used to discuss the use of modern transport technologies

4

• “Smart City” is furthermore used to discuss the use of modern transport technologies

(f) in everyday urban life (Komninos, 2008; Hollands, 2008; Alkandari et al., 2012)

� Intelligent transport systems/Infomobility contribute to the rational exploitation of existing

infrastructure without resorting to the establishment of new facilities: 1) improve the

economic productivity of current and future systems; 2) environmental protection; 3) improve

the level of traffic safety; 4) increase the prosperity of travelers, commuters and residents; 5)

increase the operational efficiency of the transportation system; 6) reduce commuting time

and cost; 7) predict the movement of traffic and events that may affect the future

• These six features connect with traditional regional and theories of urban growth and

development

Literature review of Smart Cities (2/5)

Determinants of “Smart Cities” in the literature

Urban

characteristics

Indicators

Smart economy

R&D expenditure; Employment rate in knowledge-intensive sectors; New businesses registered;GDP per employed person; Unemployment rate; % of employed in providing ICT services andproducts; etc

Smart people

Top research centres, top universities; Population qualified at levels 5-6 ISCED; Share of peopleworking in creative industries; etc.

5

people working in creative industries; etc.

Smart governance

Expenditure of the municipal per resident; Availability of new channels of communication for thecitizens (e.g. eGovernment, eHealth, etc.); Satisfaction with quality of public and social services;etc.

Smart environment

Accumulated ozone concentration; Green space share; Efficient use of water, Efficient use of electricity; etc

Smart living

Museums visits per inhabitant; Theatre attendance per inhabitant; Satisfaction with quality ofhealth system; Importance as tourist location; Overnights per year per resident; Poverty rate; etc.

Smart mobility

Public transport network per inhabitant; Broadband internet access in households; Traffic safety; Availability of ICT and modern and sustainable transport systems; etc.

Literature review of Smart Cities (3/5)

• In sum, the application of Intelligent transport systems/Infomobility in “Smart Cities”

can produce various benefits (Harrison and Donnely, 2010)

� Reducing resource consumption, notably energy and water, hence contributing to reductionsin CO2 emissions

� Improving the utilization of existing infrastructure capacity, hence improving quality of lifeand reducing the need for traditional construction projects

Making new services available to citizens, commuters and travelers, such as real-time

6

� Making new services available to citizens, commuters and travelers, such as real-timeguidance on how best to exploit multiple transportation modalities

� Improving commercial enterprises through the publication of real-time data on the operationof city services

� Revealing how demands for energy, water and transportation peak at a city scale so that citymanagers can collaborate to smooth these peaks and to improve resilience

� Drivers receive better information about traffic and road conditions and make decisionsabout which routes to follow

Health end-point Units (per year) EU25 ItalyMortality – life expectancy reduction

Months 8.6 9.0

Mortality – long term exposure Life years lost x1,000 3618 498Mortality – long term exposure Number of premature

deaths x1,000348 51

Infant mortality Cases x1,000 0.6 0.08Chronic bronchitis Cases x 1000 163 24Respiratory hospital cases x 1000 62 9

Literature review: traffic pollution and health (4/5)

Traffic

pollution still

harmful to

health in

many parts of

Europe.

7

Respiratory hospital admissions

cases x 1000 62 9

Cardiac hospital admissions Cases x 1000 38 5Restricted activity days Days x 1000 347687 48105Respiratory medication use (children)

Days x 1000 4218 531

Respiratory medication use (adults)

Days x 1000 27742 4003

Lower respir. symptoms (children)

Days x 1000 192756 21945

Lower respir. symptoms in adults with chronic disease

Days x 1000 285345 40548

Transport in

Europe is

responsible for

damaging

levels of air

pollutants and

a quarter of EU

greenhouse gas

emissions.

Source:CAFE 2005

Literature review: how intelligent transport systems can

reduce pollution (5/5)

2) providing real time Information about air pollution to the public

- spontaneous changes in mobility behavior

1) infomobility ���� easier use of public transport ���� changes in mobility behavior

���� reduction of urban pollution

Three main channels:

8

- spontaneous changes in mobility behavior

- traffic restrictions from local autorities

3) speed control traffic signals

- Kan, A. and de Barros, A.G., (2007) “The role of intelligent transport systems in reducing the

impact of traffic pollution on the environment and health”

- Bell, M. C. (2006). Environmental factors in intelligent transportation systems. IEE Proceedings:

Intelligent Transportation Systems, 153(2), 113-128.

- Coelho, M. C., Farias, T. L., & Rouphail, N. M. (2005). Impact of speed control traffic signals on

pollutant emissions. Transportation Research, Part D (Transport and Environment),10(4), 323-40.

Aim of the paper

• Drawing on Smart City’s framework, we aim to develop a model to explore the

influence of some urban characteristics of “ Smart Cities ”””” on the

provision/diffusion of Infomobility services

9

• We aim to apply this framework to 140 European cities, employing an unusually

detailed and statistically consistent dataset on public e-services at the city-level

• We analyse the relationships between urban pollution and ITS development

Data collection (1/3)

1) Urban Audit Dataset (source: Eurostat)

• Aim: providing reliable information, comparable amongst 322 cities in 27 Member States, plus 47 citiesfrom Switzerland, Norway, Croatia and Turkey

• Sample design: cities were chosen on the basis of the following criteria:

� the selected cities in each country should correspond to approximately 20% of the nationalpopulation

� the participating cities in each country should represent about 20% of the population in thatcountry

10

country

� the participating cities should reflect a good geographic distribution within the country (peripheral,central)

� coverage should reflect a sufficient number of medium-sized cities (medium-sized cities having apopulation of 50000 – 250000 inhabitants, large cities with >250 000)

Time coverage: five waves

� 1989 - 1993; 1994 - 1998; 1999 - 2002; 2003 - 2006; 2007 – 2009

Variables: nine different areas of variables have been defined

� demography, social aspects, economic aspects, civic involvement, training and education,environment, travel and transport, information society, culture and recreation

Data collection (2/3)

2) EIBURS-TAIPS Dataset (source: University of Urbino)

• Aim: desk analysis conducted through website-surfing to monitor public e-serviceavailability provided by local public transport companies and municipalities at the citylevel (EU-15)

• Sample design: 229 cities composing the EU15 subsample of the 322 (EU-27)

monitored in Eurostat’s Urban Audit dataset

Time coverage: 2012

11

• Time coverage: 2012

• Variables: two service categories have been considered, and data have been collectedadapting and integrating extant methodologies

� ITS/Infomobility (based on ITIC-Between methodology, 2010)

� eProcurement (based on IDC methodology, 2010)

Data collection (3/3)

• ITS/Infomobility: service list

Unit of analysis Local public transport company

Public Informed Mobility Electronic services related to public transportation (bus, metro, trains, etc.)

Online info to users while travellingPublic transport companies providing online information to users (e.g. waiting times,

strikes, delays, failures, etc.)

Online time table consultationPublic transport companies offering the possibility to consult the online timetable of

public transport network

12

Service

list Online travel planning

Public transport companies offering timetables with route planning (travel planner) on the web

Online ticket purchase Public transport companies offering web based payment systems

Private Informed Mobility Electronic services related to private transportation (cars, trucks, etc.)

Info to car drivers while travellingPublic transport companies providing online information to travelers about traffic or

parking

Electronic road or parking toll Public transport companies offering a electronic ticketing system of parking spaces

The construction of Infomobility Composite Indicator (ICI)

• The framework is based on ITIC-Between, 2010 and composed of 4

basic indicators

Basic indicatorService involved

(see slide 12)Variable

No. of channels used to offer information services to public transport users while travelling (call center, SMS, website, etc.)

Online info to users while travelling info_users

13

No. of different ways to access to time tables of public transportation (download, static webpage, travel planner offered via website, smart phone application, etc.)

Online time table consultation

Online travel planningtimetables

No. of different ways to purchase the ticket (smart card, website, mobile phone, etc.)

Online ticket purchase tickets

No. of channels used to offer travel info on parking and traffic to car drivers (call center, SMS, website, etc.)

Info to car drivers while travelling travel_info

Note: the service “Electronic road or parking toll” is not included in the CI since its variance is close to zero

The construction of Infomobility Composite Indicator (ICI)

• The methodology for computing the index is based on the JRC-OECD Manual for

constructing composite indicators (OECD, 2008. pag. 89)

• The weights are obtained through a Nonlinear Principal Component Analysis, which is

suitable for qualitative variables. See Gifi A. (1990) Nonlinear Multivariate Analysis.

John Wiley & Sons

Dimensions revealed COMPONENTS LOADINGS from non-

14

Dimension

Variance Accounted For

Total (Eigenvalue)

% of Variance

1 2.871 71.7682 .747 18.6653 .310 7.7494 .073 1.818Total 4.000 100.000

Dimensions revealed

Dimensions

weight1 2

tickets 0.13 0.82 0.21

Info_users 0.30 0.09 0.27

timetables 0.27 0.00 0.25

travel_info 0.30 0.08 0.27

Sum 1 1 1.00

COMPONENTS LOADINGS from non-

linear PCA SQUARED & weights

The final index is obtained as the weighted mean of the values of the 4 indicators

The diffusion of the Infomobility Composite Indicator (ICI)

Values of the CI in the selected cities (normalized MIN-MAX)

0,6

0,7

0,8

0,9

1

15

0

0,1

0,2

0,3

0,4

0,5

Fran

kfu

rt a

m M

ain

Kø

be

nh

avn

Leip

zig

Be

rlin

Öre

bro

An

twe

rpe

n

Bru

gg

e

Lin

köp

ing

Gö

teb

org

Re

ims

Ge

no

va

Fre

ibu

rg im

Bre

isg

au

Bo

nn

Tori

no

Od

en

se

Dre

sde

n

Jön

köp

ing

Han

no

ver

Ensc

he

de

Alm

ere

Dijo

n

Trie

r

Ro

ma

Nic

e

Ne

wca

stle

up

on

Tyn

e

Wo

lve

rham

pto

n

Mag

de

bu

rg

Mö

nch

en

gla

db

ach

Kö

ln (

Co

log

ne

)

Bie

lefe

ld

Wie

n

Esse

n

Wre

xham Lin

z

Mo

ntp

elli

er

Inn

sbru

ck

Lon

do

n

Pal

ma

de

Mal

lorc

a

Lee

ds

Co

rk

Gal

way

Bo

rde

aux

Am

ste

rdam

Bri

sto

l

Po

rtsm

ou

th

Sto

ke-o

n-t

ren

t

Re

ge

nsb

urg

Stra

sbo

urg

Bre

me

n

Dar

mst

adt

Re

nn

es

Du

blin

Be

lfas

t

Fire

nze

Edin

bu

rgh

Sain

t-Et

ien

ne

Len

s -

Lié

vin

Lisb

oa

Val

en

cia

Ch

arle

roi

Nam

ur

Mu

rcia

Live

rpo

ol

No

ttin

gh

am Lille

Tre

nto

Po

itie

rs

Par

is

Ro

sto

ck

Me

tz

Leic

est

er

Sale

rno

Vig

o

Kie

l

Gö

ttin

ge

n

Cle

rmo

nt-

Ferr

and

Tou

rs

Có

rdo

ba

The

ssal

on

iki

Cat

ania

Tilb

urg

Bre

da

Vo

los

Setú

bal

Cam

bri

dg

e

Lin

coln

Gro

nin

ge

n

Ko

ble

nz

Aja

ccio

Ve

ron

a

Car

dif

f

Co

imb

ra

Po

tsd

am

Mad

rid

Gijó

n

Trie

ste

Sch

we

rin

Ein

dh

ove

n

Ovi

ed

o

Irak

leio

Ho

spit

ale

t d

e L

lob

reg

at(L

')

Pal

erm

o

Cam

po

bas

so

Ave

iro

Fun

chal

Cat

anza

ro

Po

nta

De

lgad

a

Bad

ajo

z

Po

ten

za

Pe

scar

a

Re

gg

io d

i Cal

abri

a

Ioan

nin

a

Log

roñ

o

Fort

-de

-Fra

nce

Gra

vesh

am

EU15 average

The diffusion of the Infomobility Composite Indicator (ICI)

Average values of the CI in the selected cities, by Country (normalized MIN-MAX)

0,5

0,6

0,7

0,8

0,9

EU15 average

16

0

0,1

0,2

0,3

0,4

DK SE LU BE DE AT IE UK NL FI FR ES IT PT EL

The theoretical pinpoint of the analyzed model

Where Infomob is our dependent variable (composite indicator), eproc is a composite indicator of

iiiii

iiii

iiiii

carsictlocuntournightsozone

highedemplfinempltranspemplhotel

emplrpopdenspopeprocInfomob

εββββββββ

ββββα

++++++++

+++++=

1211109

8765

4321 lnln

17

Where Infomob is our dependent variable (composite indicator), eproc is a composite indicator of

eProcurement (calculated as the simple mean of 13 indicators), lnpop and lnpopdens are the

logarithms of population and population density respectively, emplr is the municipal rate of

employment, emplhotel, empltransp and emplfin are the employment in hotels-restaurants-

trade, transport-communications and financial and business sectors (in %), highed is the share of

people between 15 and 64 years of age with at least a degree (ISCED 5-6), ozone is the number of days

in a year when an excess of ozone is recorded in town, tournights is the number of tourist overnight

stays in registered accommodation per year per resident population, ictlocun is the proportion of local

units producing ICT, cars is the number of registered cars per 1,000 inhabitants.

Data are taken from the waves of Urban audit Eurostat giving priority to the last available figure.

Descriptive statistics

Max= Stockholm 2008

18

229 towns. 195 only theoretically available. Due to missing data of towns in some of these variables the

number of obs. On which the econometric analysis is made goes down to 140. Still it is a very high figure

Econometric model: results (1/4)

Positive effects are found for all of the

significant regressors even though with

different p-values

19

Adjusted R2= 0.344

P-values (+0.1 *0.05 **0.01 ***0.001)

N=140!!! It is the first time that an

analysis of this kind is performed on such

a number of towns

Econometric model: results (2/4)

• The provision of infomobility services is strongly related to the size of theEuropean cities (variable expressed in terms of total population in the city)

� external pressure on Local Public Transport Companies (LPTC) to improve services can beexpected to increase with the number of city inhabitants

� the perceived need for advanced communication tools between LPTC and citizens appears toincrease with size, hence with the physical and social distances to be covered within the territoryof the city in order to gain access to service providers

20

• Another important factor affecting the availability of Infomobility tools includethe economic structure of the European cities, with a positive correlation offirms and workers in knowledge intensive services (financial and businesssectors)

• We observe that the availability of Infomobility services is affected by thepresence of other innovative actors in the same city

� Among these actors are the municipalities offering eProcurement services

� This result proves that when a high number of innovators are located in a given area, knowledgespillovers will be facilitated and greater incentives are created that push less dynamic institutionsto enter the innovation race

Econometric model: results (3/4)

• The presence of local ICT producers in the city is also positively correlated withInfomobility development

� Local Public Transport Companies located in cities with higher shares of local ICT producers are ina better position to gain access to relevant technology, including both hardware and software

� Where public and private markets overlap, as in the case of voice or image transmission over IPand value added services to business enterprises, a competitive presence of ICT serviceproviders stimulates the public organizations to expand the range of services offered through

21

providers stimulates the public organizations to expand the range of services offered throughtheir city networks

• The level of pollution has an impact on the development of ITS (see next slides)

• Finally, we find a positive correlation of Infomobility Index with the employmentrates in the European cities

� It appears that Local Public Transport Companies that are located in dynamic areas tend tointensify their provision of e-services

� Employment rates are logically associated with the quality of social environment in which localadministrations operate and with the level and sophistication of demand for services expressedby citizens and firms

Econometric model: results (4/4)

The model is well specified (Reset test is ok) and is robust to changes in the scale of measurement (i.e. use of logs for some variables or percentage for others), homoskedasticity is verified through Breusch-Pagan test, normality of residuals through the Shapiro-Wilk test and standard graphical procedures (pnorm qnorm). Some influential city (9, through Cook D’s threshold of 4/n) are

the following:

1) Aarhus (Den, medium infomob)

2) Paris (Fra, medium infomob)

High, mediu

m, low

22

2) Paris (Fra, medium infomob)

3) Luxembourg (Lux, high infomob)

4) Aalborg (Den, high infomob)

5) Cremona (Ita, low infomob)

6) Edinburgh (UK, medium infomob)

7) Stockholm (Swe, high infomob)

8) Venice (Ita, high infomob)

9) Madrid (Esp, low infomob)

m, low

based on

percentiles

Econometric model: results for days of ozone excess (1/2)

11

.52

2.5

3D

fbet

a d

ays

_oz

on

e_ex

cess

_la

t_a

v

Influence of city on ozone standard error

(threshold 2/sqrt(n))

23

Cremona

BolognaVerona

CampobassoBadajozDarmstadtPotsdam RomaFirenze MalmöMülheim a.d.Ruhr Aix-en-ProvenceMainz Cagliari TurkuMadridToledo UtrechtBielefeld Caen TorinoNürnberg BariBarcelona ToulonDortmund StevenagePoitiers CatanzaroDüsseldorf Lens - Liév in NijmegenAalborg Palma de MallorcaReims GöteborgAnconaAmiensBremen Zaragoza Besançon WienToulouseHannover LyonMönchengladbachLogroñoRennes BirminghamLilleKøbenhavn HeerlenRouenSantanderNancyGöttingen LimogesBordeauxEssen Stoke-on-trentSaarbruckenAugsburg PortsmouthPointe-à-PitreOdense RegensburgHamburg Metz HelsinkiParisKöln AjaccioNapoliErfurt Saint-EtienneCataniaStuttgart GrenobleCayenneMálaga ExeterGroningenStrasbourgRostock Le HavreDijonOrléansClermont-FerrandBochum ToursBruxelles / Brussel NantesPamplona/IruñaValenciaMurcia PerugiaMarseilleKiel Trento LiverpoolMontpellierLeipzig ManchesterL'AquilaPescaraRotterdamAmsterdamDresdenSchwerin BelfastNice BredaMagdeburgMoersBerlin Bonn PalermoSaint DenisTriesteSevilla Fort-de-FranceTrierMünchen VeneziaAarhusFrankfurt am Main 's -GravenhageLuxembourg (city)PotenzaKoblenz EdinburghHalle an der SaaleWeimar MilanoFrankfurt (Oder)Karlsruhe StockholmGenovaFreiburg im BreisgauWiesbaden

-1-.

50

.51

Dfb

eta

da

ys_

ozo

ne_

exce

ss_

lat_

av

0 50 100 150 200 250Id

(threshold 2/sqrt(n))

Italian and German cities respectively

lower and increase the coefficient of

the pollution variable by a strong

amount

Econometric model: results for days of ozone excess (2/2)

Ca enBielef eld

Ma lm ö

Stevena geP oitiersDortm un dDüsse ldo rf

M ülh eim a.d .Ru hr

Turku

B arcelon aZarag ozaB esan çonGöte bo rg

ReimsAalb org

Købe nh avn

Roue n

Od enseErfurtMála ga

Dijo nBru xelles / Brussel

Le ipzig

S trasbo urg

M oersBo nn

Berlin

M arseille

Trie r

Dre sde n

Fran kfurt a m Ma in

L uxembo urg (city)

Ma gd ebu rgM o ntp ellie r

Fran kfu rt (Od er)

Ve ne zia

Halle an d er Saa leW eimar

W iesb aden

Ka rlsruh eGen ova

Sev illaNice

Freibu rg im Breisga u

M ilano

M urcia

.51

e(

info

mo

b_

n | X

)The same story.

Italian and German

cities are influential

on the pollution

effect on

infomobility

24

Pote nza

Cag lia ri

L'Aq uila

Cata nzaroBariAn con a

Roma

Ko blen z

Peru gia

Pa lma de Ma llo rca

Sto ckho lm

Mün chenPale rm o

Na poli

Nü rn be rg

Edinb urghTrie ste

Va lencia

Ca en

Belfast

Reg en sburg

Bielef eld

S antanderS toke-o n-tre nt

C le rm on t-Fe rran d

Stevena geLe ns - Liév inP oitiers

L imo ge s

Pamplo na /Iruña

Schwerin

Nant esMa nch este rL iverpo ol

Renne s

Ro tte rd am

Birmin gh am

's-Grave nh ag e

B ordea ux

Dortm un d

Rostoc k

Düsse ldo rf

B re daAmst erdam

Aarh us

Orléa ns

Saint De nis

Kiel

S aint-Etie nneHe lsin ki

Fort-de -Fran ce

Han nove rM ön che ng la db ach

Exe ter

Tours

Zarag ozaB esan çon

B ochu m

Esse nHam burg

Nan cy

Pe scara

Greno bleA ja ccio

Le HavreGro nin gen

Köln

P ointe -à-Pitre

Ca ye nn e

Od ense

Metz

Pa ris

Erfurt

Sa arbrucken

L ogro ño

Mála ga

Lyon

Tou louse

L ille

Nijme ge n

Göttin ge nAug sbu rg

Amie ns

S trasbo urg

Portsmou th

M ad ridM ainz

To ulon

W ien

M arseille

He erle n

Stuttg art

Brem en

Aix-e n-Pro ven ce

UtrechtToled o

Po tsda m

M o ntp ellie r

Firen ze

Tre ntoCa ta nia

Sev illa

Darmsta dtB ada joz

Nice

V erona

M urcia

Crem on a

To rino

B olog na

Cam po ba ss o

-.5

0e

( in

fom

ob

_n

| X )

- 40 -20 0 20 40e( days_ozone_ex cess_la t_av | X )

coef = .00 32 48 15, se = .0 01 697 89 , t = 1.91

Grouping Analysis (1/4)

The nations of the 140 towns in the regressionThe nations of the 54 towns in the regression belonging to

the group with high pollution (days of ozone excess)

25

The nations of the 27 towns in the regression belonging to the

group with high pollution (days of ozone excess) and high

infomobilityMore than 70% of the german cities

with high pollution developed a high

level of infomobility, while the same is

true for less than 40% of the Italian

cities and 30% of French cities with

high pollution.

Grouping Analysis (2/4)

High pollution - High infomobility High pollution – Low infomobility

26

Towns with High pollution and High infomobility show in average an higher level

of eProcurement, an higher level of employment rate and employment in the

financial and business sectors. They are also slightly bigger. Town with High

pollution and low infomobility are in average more polluted.

Grouping Analysis (3/4)

High pollution - High infomobility

27

Grouping Analysis (4/4)

High pollution - Low infomobility

28

Conclusion

• There is a significant heterogeneity in the infomobility diffusion betweenEuropean cities reflecting demand-pull considerations

• We showed that innovative activities of Local Public Transport Companies(LPTC) also reflect interdependencies among a variety of actors, especiallythose active in the same city (municipalities and local ICT producers)

• There are important contextual factors which complement demand andsupply factors as key drivers for innovation in the Infomobility services

• German cities are very widely represented among those belonging to the

29

• German cities are very widely represented among those belonging to the“high infomobility-high pollution” group (15 out of 21), while Italian (andFrench) cities are much less so. More than 70% of the german cities withhigh pollution developed a high level of infomobility, while the same is truefor less than 40% of the Italian cities and 30% of French cities with highpollution � national variables matter

• This results illustrates that in the latter cases (Italy and France) infomobilityis carried out largely regardless of the actual need of cities to reducepollution. This might indicate that in many circumstances infomobilitypolicies are designed more at the national than at the local level, and hardlyreflect actual priority of municipalities to control pollution levels.

Thanks

30

Data collection

• eProcurement: service list

Category eProcurement MunicipalityUnit of

analysis

eProcurement Visibility Measures whether the municipality make available eProcurement services to potential suppliers on their web site

Publication of general information on public

procurementGeneral information about the public procurement made available on the municipality websites

Publication of notices to official electronic notice boardsAvailability of an official electronic notice board on the municipality websites where the procurement

notices are made publicly available Link to e-procurement services Availability of a link to a web page providing eProcurement services. The web page may be part of the

website owned by the municipality or part of the website owned by an external supplier

eProcurement (Pre-Award Phase)Measures the availability of 3 sub-phases (e-NOTIFICATION, e-SUBMISSION, e-AWARDS) constituting

the eProcurement process

e-NOTIFICATION Publication of tenders and procurement notices on the web

Online registration of supplier Creation of user accounts and profiles with related roles

e-mail alerts for suppliers Possibility for the suppliers to receive email alerts about forthcoming calls and notices of their interests

31

Service

list Service

description

e-mail alerts for suppliers Possibility for the suppliers to receive email alerts about forthcoming calls and notices of their interests

e-SUBMISSION Submission of proposals online

Assistance services to the supplierAvailability of online communication channels (e-mail, chat, audio/videoconferencing) to carry out Q&A

(Question and Answer) sessions between the eProcurement operator and the bidders

Online supplier help sessionExistence of specific user help services, finalized to the assistance of the supplier for the preparation of

the online tender

e-AWARDS Includes the publication of awarded contracts

Online information about awarded contracts The website publishes the contracts awarded and their winner

e-auctions Availability of tools to carry out real-time price competitions

eProcurement (Post-Award Phase)The eProcurement Post-Award Process measures the availability of 3 distinct steps (e-ORDERING, e-

INVOINCING, e-PAYMENT) constituting the procurement process after the award of the contract

e-ORDERING Automatic placement of orders online

e-cataloguesPossibility to order online from e-catalogues managed by the eProcurement website and structured

according to the type of procurement, the product/services prices and characteristics

Electronic marketAvailability of an electronic market hosted by the eProcurement website, for the online interaction

between buyers and suppliers

e-INVOICING Delivery of electronic invoices

e-invoicing service Availability of e-invoicing services managed by the eProcurement website

e-PAYMENT Online payment of contracts

e-payment service Availability of online payment services, managed by the eProcurement website

Econometric model: post-estimation diagnostics (1/2)

Max vif: empl hotel 3.19

Mean vif: 1.87

Threshold vif: 5

Breusch-Pagan test: chi-squared (1 dof)

P-value 0.2114

Normality: SW test = -0.406 P-value 0.658

Specification RESET test =1.14 P-value 0.3366

32

Specification RESET test =1.14 P-value 0.3366

Four light outliers (studentized res >|2|) not exceeding iqr range:

1) Cremona (Ita) minus (low infomob)

2) Paris (Fra) minus (medium infomob)

3) Mainz (Ger) minus (low infomob)

4) Wiesbaden (Ger) plus (high infomob)

Seven possible leverage points >(2k+2)/n:

1) Aarhus (Den, medium infomob)

2) Luxembourg (Lux, high infomob)

3) Venice (Ita, high infomob)

4) Edinburgh (UK, medium infomob)

5) Palma de Mallorca (Esp, medium infomob)

6) Rome (Ita, high infomob)

7) Stockholm (Swe, high infomob)

Econometric model: post-estimation diagnostics (2/2)

Cam pobassoTorino P arisAa lborgStoc kholmRoma

Palma de Mallorca

Edinbu rghVenezia

Luxembourg (city)

A arhus

.2.3

.4L

eve

rag

e

High leverage

33

Metz Reim sStrasbourg Rouen Ma inzTou rsAugsburgRostockNancy DijonK arlsruheCaenKiel TrierBielef eldErfu rtBochumNantes LyonMarse illeBremenBordeauxHannoverLilleM agdeburgStutt ga rtKölnClermont -Fe rrandDresden GroningenRegensburgBes ançon 's-GravenhageLe ipzigAmiens B onnDüsse ldo rfOrléansMönchengladbachEssen DortmundRennesSaint -Et ienneHeerlenSchwerinW ienLive rpoo lBirmingham LogroñoGrenoble Koblenz Po tsdam W iesbadenNijmegenHalle an der SaaleMa lmöNice ZaragozaMontpellierNürnberg MoersTou louseDarmstad tRo tterdamL imogesA ncona Mü lheim a.d.Ruh rPoitie rsTren toA jaccio Tou lonTriesteVa lenciaBe lf ast ToledoMálagaS ant anderHam burgCatan iaStoke-on -trent Fo rt -de-FranceLens - LiévinMünchen Fre iburg im Bre isgauVerona Frankfurt (Ode r)A msterdamMancheste r Be rlinK øbenhavnPerugiaExete rPorts mouth CremonaS evillaGöttingen MadridPescaraW eimarPale rmoStevenage Frankfurt am M ainAix-en-ProvenceUt rechtBredaMilano BarcelonaMurcia Sain t Den isGö teborgBari Genova TurkuP ointe -à-Pitre B ruxe lles / B russe lLe HavreS aarb ruckenNapo liCag lia riPamplona /IruñaHelsinki FirenzePo tenzaCatanzaro BolognaL 'A quila B ada jozOdenseCayenne

Cam pobassoTorino P arisAa lborg

0.1

Lev

era

ge

0 .01 .02 .03 .04Norm alized residual squared

High residual