Modeling and simulation of Cohesion Policy
funding and regional growth diffusion in an
enlarged European Union
Business School of Normandy, Métis Lab,
9, rue Claude Bloch, 14052 Caen (France)
Abstract : The dilemma between equity and competitiveness has created con-
cerns about the future of redistribution of European regional policy funding.
The objective of this chapter is to estimate the spatial expression of conver-
gence and regional growth in the European Union. After contextualizing the
EU enlargements of 2004 and 2007, this study uses spatial statistics and the
simulation platform GeoCells, the goal of which is to analyze two alternatives
for future economic development of the EU.
Keywords: cellular automata, cohesion policy, regional disparities, equity and
Acknowledgements : This work is part of a collaborative project led with Bernard
Elissalde, Dominique Goyat and Patrice Langlois, all of whom we wish to thank
On May 1st, 2004, the most important extension of the European Union (EU) in
history took place. Ten countries became full EU members: in the north, the three
Baltic States (Estonia, Latvia and Lithuania), the four countries of Central Europe
(Hungary, Poland, the Czech Republic and Slovakia), a country of south-west area
(Slovenia) and two islands (Cyprus and Malta). Two countries in South-East Europe
(Bulgaria and Romania) integrated the EU on January 1st, 2007. Consequently, the
level of prosperity in the EU declined significantly. However, because of the long
process of transformation of post-soviet societies, this event was generally received
Numerous geographical issues arose from this policy of openness in the Central
and Eastern European Countries (CEECs). What territorialized management of the
cohesion policy was required with the arrival of ten new countries? The community
economic frame was disrupted by the last two enlargements which provoked an
unprecedented increase in the economic gap between developed regions and those
lagging behind. This situation requires the member states to revise the objectives
regarding cohesion in order to prevent increasing economic, social and territorial
fragmentation of the Union.
The inclusion of the CEECs, countries with far less economic development than
the poorest of the EU-15 (Italy, Spain, Greece and Portugal) reopened the question
of the ability of Europe to promote socioeconomic and territorial cohesion. In light
of the results of our simulations, our prospective approach proposes two possible
scenarios of economic development for the EU of tomorrow by demonstrating the
dilemma between equity and competitiveness (Lackenbauer 2006).
The purpose of this research is to understand the process of convergence by
using the simulation platform GeoCells (Elissalde et al., 2009) coupled with spatial
statistics. An application of this model demonstrates the economic performance of
European regions according to the variation in aid granted by the European Union,
as well as neighborhood effects. Taking into account the regional disparities,
GeoCells analyzes European regions’ relative positions from the angle of
macroeconomic and budgetary indicators. The cellular automaton GeoCells allows
an assessment of the overall effectiveness of regional policy and measures the
influence of modification of granting rules.
The introduction of simulation and forecasting methods, along with spatial
statistics, in EU regional policy debates is not an attempt to find the one and only
response to the problem of European regions’ unequal development. Instead, it
suggests a range of credible options as a decision support tool for territorial
solidarity – as well as economic and social cohesion – in a European space which
is in perpetual evolution. Even though European regions belong to an
interdependent group, they each have their specific trajectories, in which reaction
times and pace of change vary strongly from one to another. These various
trajectories build a European regional mosaic, making it difficult for policy makers
to override initially planned regional policies (Cohesion Policy, Cohesion Funds,
etc.) with budgetary adjustments. Overarching policies are enacted for these
separate states/regions in their separate trajectories - but these policies may actually
prevent, curtail, or disproportionately power certain trajectories, and in fact may
disable newer "corrective" policy/fiscal mechanisms from assisting.
Methodologically, the GeoCells cellular automaton is based upon interactions
between variables (e.g. time periods, growth rates in the GDP per head, flows of
public investments) and three geographical levels (European level, national level
and regional level). Due to the role of spatial interactions and contiguity effects in
regional trajectories, in this research, a regional growth diffusion parameter was
added to the above variables ratified by the European Commission. Though many
regional growth models analyze the region as a stand-alone unit and ignore spatial
interaction phenomena linked to proximity, neighborhood, or contiguity effects, the
spatial dynamic parameter was added to the variables to underline the role of growth
diffusion in regional development.
2. What is the role of European regional policy in reducing
disparities in the EU?
The issue of the solidarity effort between Member States and regions (NUTS 2),
as well as their adherence to the cohesion principles defined in the European texts
and treaties, is at the center of the debates on European regional policy. The
European Union's regional policy seeks to reduce structural disparities between EU
regions, foster balanced development throughout the EU and promote real equal
opportunities for all. Based on the concepts of solidarity and economic and social
cohesion, it achieves this in practical terms by means of a variety of financing
operations, principally through the Cohesion Policy (European Regional
Development Fund (ERDF) and the Cohesion Fund). For the period 2007-2013, the
European Union's regional policy is the EU's second largest budget item, with an
allocation of EUR 348 billion. The objective of economic and social cohesion was
introduced in 1986 with the adoption of the Single European Act. The policy was
finally incorporated into the EC Treaty itself (Articles 158 to 162) with the
Maastricht Treaty (1992).
The main question is in regard to the ability of Cohesion Policy to reduce
disparities produced by the single market. How can we improve redistribution and
territorial equity in a Union with low economic growth? In such an economic
context, should we limit the solidarity efforts of wealthy countries or, on the
contrary, increase it in order to accelerate the economic advancement of regions in
an earlier stage of economic development?
The implicit deal between the EU and CEECs of opening new markets against
the backdrop of the promise of a rising standard of living for relevant populations
also implies that this development is achieved by offering newcomers Cohesion
Policy. The results of EU policies in helping regions to economically advance are
very difficult to assess accurately.
The evaluation of effectiveness of Cohesion Policy in promoting regional
development raises methodological problems (Fayolle and Lecuyer 2000). Even if
the distribution of Cohesion Policy is proportional to the economic development
level, and regions lagging behind are catching up with wealthier regions, it is
difficult to determine whether these outcomes are due to Cohesion Policy or other
factors. In addition, there is no guarantee that the Cohesion Policy constitutes an
explanatory factor of the regional convergence, even though their correlation is
significant (ESPON Project 2.2.1. 2005). Indeed, we cannot rule out the possibility
that a natural convergence process is simply an outcome of developmental progress.
Following the integration of ten CEECs in the EU in 2004, a debate on the
development of the poorest regions emerged in the mid-2000s. The European
Commission (2006, 2008a) hoped to invest massive resources in order to help them
to develop more quickly. Nevertheless, Gorzelak et al. (2010) argues that the
development through a massive injection of money in poor regions is ineffective.
The transfer of more than EUR 1 billion euros did not meet expectations in southern
Italy and former East Germany (ibid.). In addition, this method of massive
investment has a perverse effect: it can create a situation in which the inhabitants of
these regions become dependent upon aid they receive. Aghion and Cohen (2004)
have shown that the only effective regional investment for poor regions is
investment in education. However, these Funds are traditionally invested by the new
Member Countries (a decision-making power which the EU has allowed) in other
infrastructures such as transport. It is therefore understandable that policies in
southern Italy or in Extremadura have not been fruitful.
The EU has to face to another structural obstacle. The EU must accept that the
regional disparities in Eastern Europe have existed for centuries. It is very difficult
to change these disparities in the time frame outlined in the Cohesion Policy
program (i.e., 2000-2006 or 2007-2013). In Poland, for example, Coudroy de Lille
(2009) highlighted the fact that regional contrasts and their spatial inscription were
created in the nineteenth century. Stryjakiewicz (2007) also explains that
metropolization has accentuated the regional disparities during economic transition.
Thus, the CEECs are fragmented within their own borders, with disparities between
cities and the countryside and between West and East. These differences are
reinforced in the historical distribution of wealth, the post-Soviet transformation,
the values of society and the efficiency of government.
Finally, before making decisions about fund allocation, it is necessary to
consider where to invest. One might think that for ethical reasons, that aid should
go to the poorest regions. However, studies show that investing in cities has much
more of an impact than investing in rural communities (European Commission,
2008b). The analysis of successive generations of European aid to CEECs
highlights the dilemma between equity (investment in rural areas with the goal of
convergence) and competitiveness (investment in cities with the goal of growth).
The economist Williamson (1965) studied the contradiction between a strong GDP
growth rate and the increase of regional disparities. These studies were recently
replicated in the EU by Ezcurra and Rapun (2006), who also came to the conclusion
that an increase in financial support for CEECs would produce simultaneous
convergences between the growth rates of CEECs and member countries of the EU,
while increasing regional disparities within the CEECs. According to Bergs (2001)
inter-regional convergence could take place over time, but at the expense of the
national growth potential of new members. The latest report from the European
Commission on Economic and Social Cohesion seems to confirm this prediction. If
the disparities in the GDP per capita are decreasing between countries, they are
increasing in each country. This is the case for both EU-15 Member States and the
new Members States (European Commission, 2006 and 2008a; European
Parliament, 2007 and 2008). Thus the problem of competitiveness and equity is
posed (Fayolle and Lecuyer, 2000): should we help the least developed regions in
order to help them to catch up?
Although the EU structural policy remains an important instrument of cohesion
and solidarity at European level, its effectiveness at the EU regional policy level
needs to be considered. However, because of the myriad of factors that come into
play, it is impossible to assess categorically the true impact of the Cohesion Policy
on European Spatial Planning (Durh et al., 2009) and territorial cohesion (Jouen,
2008; Kilper, 2009). It is also difficult to know what beneficiary regions would look
like today if the funds had not been granted. It is for this reason that the modeling
and simulation of EU Cohesion Policy based upon the configuration of regional
economic disparities could contribute to the evaluation of european policies.
3. Toward modeling the cohesion policy and its effects upon
regional economic dynamics
With the aim of investigating possible solutions for reducing the development
gap – a gap which increased significantly with the progressive transition from 15 to
27 Member States in the European Union – we have developed a cellular automaton.
The simulation platform GeoCells is use to determine under which conditions (in
terms of budgetary redistribution settings) and according to which goals (of
reduction, convergence, or adjustment), European solidarity policies could be
3.1. The need for modeling and simulation to understand the
issues of European regional policy
Economic theory has various tools for clarifying and analyzing the issue of the
role of European cohesion policy in the convergence process:
i) growth theories allow for an analysis of the mechanisms of economic growth as
well as the outlook for divergence or convergence of economies;
ii) theories on geographic economy allow for a study of agglomeration mechanisms
in economic activity and the spatial structure of economic disparities;
iii) econometric methods present tools for an evaluation of convergence phenomena
in conjunction with cohesion policy.
With the development of simulation methods, several macroeconomic models
have been created in order to understand the role of European regional policy in
reducing regional disparities. Such simulations allow for an evaluation of what
would have been the current situation of GDP in the absence of cohesion policy.
These models also permit ex-ante or ex-post analyses and offer scenarios according
to budgetary stance. Such a model undeniably has certain benefits. It is mainly for
these reasons that the European Commission bases itself on work carried out within
the framework of the HERMIN (Bradley et al., 1995, 2003 and 2007) and QUEST
(Roeger et in 't Veld, 1997 et 2004 ; Varga et in 't Veld, 2011) models in its European
Funds assessment reports. It draws some rather flattering estimations on the role of
regional European policy in short term growth (Kelber, 2010) for the HERMIN
model, whereas the QUEST model makes some slight references to its long term
impact (Magnier, 2004).
Several publications have, however, highlighted their limitations. For Sjef
Ederveen et al. (2002, 2006), the application of models such as QUEST et HERMIN
only gives a glimpse of the potential effects of cohesion policy in the sense that
these Funds have numerous parameters of efficiency. Nevertheless, according to
these same authors, regional policy appears to be more successful in a environment
which is conducive to growth. The example of the "Irish Miracle" is a clear illustra-
tion of this. Furthermore, Philippine Cour and Laurence Nayman (1999) note that
the simulations only assess what the economic situation would have been in the
absence of European regional policy (see for example "Panorama Inforegio", n°33,
2010) in a short term analysis. Finally, numerous underlying assumptions are made
and their generality is problematical (Cappelen et al., 2003). For example, it is taken
for granted that the collected Funds are systematically allocated to productive public
investments, an assumption which is far from being systematically verified. The
HERMIN model is based on the assumption that States are open economies (Brad-
ley, 2002), which is not the case everywhere in Europe. One of the major limitations
of this macroeconomic model is that it can only be applied on a national level. The
regional declination is overlooked in this model due to insufficiently comprehensive
databases. Moreover, amongst the assessments of the role of cohesion policy in re-
gional growth and convergence, a number of authors (Le Gallo, 2004 ; Rey and
Janikas, 2005 ; Ertur and Le Gallo, 2008) have demonstrated the role of the effects
of neighborhood and spatial dependency on the efficiency of European Funds. The
effects of diffusion of regional growth have not been taken into account in either
In this context, we consider that modeling by cellular automaton (Hill, 1993)
enables a clarification of the issues of convergence and European regional integra-
tion. The model that we have developed allows for the effects of neighborhood and
diffusion of regional growth to be taken into account. In addition, modeling by sim-
ulation is useful in that it reveals the processes and mechanisms (i) and serves as a
decision support tool (ii).
(i) Cellular automaton simulation is constructive as it takes into account the com-
plexity of the relationship between decision making (budgetary stance, duration of
European regional policy programming periods), economic factors (growth and
convergence) and spatial aspects (interaction between regions/Member States)
(ii) Simulation is helpful when it is not a question of finding the optimal solution
but of exploring a wide range of possible scenarios in order to identify the parame-
ters that would significantly improve the efficiency of European cohesion policy.
3.2. GeoCells, a multi-layered hierarchical automaton
GeoCells is a simulation platform based upon layers of geographic information.
Its main engine is a meta-model based upon spatial agents or a topologic cellular
agent. GeoCells is used to model the evolution of GDP per capita in the EU-27, and
the simultaneous influence of different types of aid under the cohesion policy, and
the effects of growth diffusion by neighborhood. The general operating principles
for GeoCells are displayed in Figure 1.
The system is based upon a group of geographic information layers (Fig. 2).
Each layer (EU (1); member-state (2); region (3)) consists of features from the same
class. Each layer is made up of cells (EU, countries, regions). A cell’s main function
is to own, in addition to the feature’s physical components (location, shape, size…),
the knowledge of its neighborhood and above all a behavior dynamic.
Each layer owns behavior rules giving to the cells of its class the same function
in the system (region, member-state, EU), properties and attributes (perimeter,
surface, budget of the cell) and relations with cells from other layers of the system.
The system takes into account the hierarchical relationships existing between
layers (Fig. 2); a region (Layer 1) belongs to a country (Layer 2) – inclusion link –
and a country is made up of regions – containing link.
Figure 1. GeoCells functioning principle
Figure 2. Hierarchy of cellular layers
The generated automaton characteristics are described below:
Cellular interactions: A cell layer interacts "naturally" with its sister cells (its
topological neighbors) with its mother cell (above in the hierarchy) or its daughter
cells (below in the hierarchy), but can also interact with cells of any other layer,
through explicit links. These links represent cellular exchanges.
State: Each cell’s state represents an attribute that is likely to change during the
simulation. Each state has a semantics which can represent information (such as the
budget of its neighbor) or an amount of material or energy (such as its own wealth,
Phases of Life: One of the difficulties of this type of mechanism is to maintain the
temporal coherence between all cellular layers. Every cell performs four steps when
it receives inflow:
- Reading of its inputs (inflow from outside);
- Implementation of its program of action (behavior);
- Writing of its outputs (outflow exchange);
- Storage of its context (each cell must maintain at least the contents of the
previous context). The context is defined here as the previous state of the cell and
the recording of the state variables of neighboring cells.
Capacity - Every cell has its attributes (or state variables) but the rules of behavior
are collective (because they are shared by all elements of its class). Each cell
generates actions that depend upon its inputs and its state at a given time. The action
taken is the result of a choice of the cell. This choice depends on the evaluation of
the relevance of the rules of actions that may apply. In other words, the cell can have
“smart” behavior comparable to that of an agent (we nevertheless retain the term
Communication canals - A bidirectional communication canal exists to combine the
system’s multilayer nature. Each cell owns the input and output references relating
to the canals that concern it. For this reason, the cell knows its environment and
enters into dialogue with it.
3.3. The possible simulation settings
Given the data available for the group of regions NUTS2 of the EU-27, the
model generated, as the main indicator, the variation in GDP per capita of each
European region. Within Geocells, policy variables are adjusted for each simulation,
while population remains constant. A user interface provides an opportunity at the
beginning of the simulation for the user to enter a value for each policy variable.
The settings which can be varied within Geocells are described below.
The Article 160 of the Treaty establishing the European Community (in its
consolidated version in 2002) provides that the European Regional Development
Fund (ERDF) is intended to help to redress the main regional imbalances in the
Community. The ERDF therefore contributes to reducing the gap between the levels
of development of the various regions and the extent to which the least favoured
regions, including rural and urban areas, declining industrial regions, areas with a
geographical or natural handicap, such as islands, mountainous areas, sparsely
populated areas and border regions, are lagging behind. Rules of allocation of
Cohesion Policy as defined in the Treaty have been implemented in Geocells. The
GDP variation rate is, either specific to the region or identical to the group of
regions of the same country or identical for the whole of EU. The terms of public
intervention include the mechanisms relating to contributions (Countries and EU),
to the aid linked to regional policy, such as eligibility thresholds (75% of the
average GDP per capita of the EU) for Cohesion Policy. The European budget
weight is taken into account. The EU budget is stabilized around a threshold of 1%
of the total European GDP (threshold reached since 1984 with the Single European
Act). The EU had an agreed budget of EUR120.7 billion for the year 2007 and EUR
864.3 billion for the period 2007–2013, representing 1.05% of the overall wealth of
the EU-27's. From this average budget, simulations were able to make the
Eurropean budget weight vary from 0,5% to 3% of the EU total GDP. The principle
of additionality between the States and the European Union in the Cohesion Policy
financing was also taken into account. According to this principle, EU funds can
only be paid in addition to a contribution from the member states, not instead of it.
The variability of the relative importance of regional policy in the EU budget
expenditures is also one of the simulation settings. The ERDF and the Cohesion
Fund make up one of the largest items of the budget of the EU. The overall budget
for the period 2007-2013 is EUR 271 billion and represent 30,4% of total EU
expenditures. In addition to these principles officially ratified by the Treaty
establishing the European Community, we have added to our model a spatial
dynamic parameter: the hypothesis of the role of spatial interactions and of
contiguity effects in the regions’ trajectories.
The diffusion by contact with neighboring regions, made possible by the
functioning of the cellular automaton, is carried out therefore naturally in one way
or another. With GeoCells, what is happening in the neighboring regions is not
ignored. Several researchers (Baumont et al. 2002, Islam 2003; Le Gallo 2004; Rey
and Janikas 2005; Dall'erba and Le Gallo 2008; Dall'erba et al. 2009; Dall’erba and
Hewings 2009; Ertur and Le Gallo 2008) have shown that most studies consider the
regions as isolated entities, as if their geographical location and their potential inter-
linkages were not important. However, the geographical distribution of growth
phenomena at the regional level is rarely random: the economic performances of
neighboring regions are often similar (Getis 1991). The impact of the unequal
distribution of economic activities in space upon the territories' economic growth
was underlined in particular by Baumont (1998). While a situation of spatial
competition between activities and between territorial units exists, the taking into
account of contagion, of mimicry phenomena linked to neighborhood effects proves
to be necessary.
3.4. Growth-diffusion model for European regions
We have attempt to model a complex diffusion process in real life by choosing
a specific diffusion mechanism. The diffusion by contact with neighboring regions
was highlighted especially by Elissalde et al. (2009) and Bourdin (2013) who has
shown for example that regions of Central Europe (eastern Germany, the western
parts of the Czech Republic, Slovakia, Hungary and Slovenia) have a low level of
GDP per capita compared to the EU15 average, but a geographic environment
which is more favorable than the regions further to the east in the EU. In this con-
text, a catching-up of regions of Central Europe is explained in part by a growth
diffusion process by neighboring. The proximity of regions of Central Europe to the
border of the EU15 gives to these regions a high development potential compared
to regions further east. This suggests that the distribution of regional growth occurs
more neighbor to neighbor.
We will now clarify the unique diffusion model that we have used. The term
Xi represents the GDP of the region i, Pi its population and Yi = Xi/Pi its GDP per
capita at a moment t.
We present the following hypothesis. Each cell has the aim to homogenize,
through time, its standard of living Y in relation to its neighbors. The attempt to
homogenize standard of living is the policy goal of the Territorial Cohesion. The
main aim of the Territorial Cohesion policy is to contribute to a balanced
distribution of economic and social resources among the European regions with the
priority on the territorial dimension. This means that resources and opportunities
should be equally distributed among the regions and their populations. But, in our
model, standard of living is not capable of diffusing like a flow. It is through the
variation of wealth (X) symbolized by the GDP (by internal growth and by
diffusion) or through the variation of population (P) (also by internal growth or by
migrations) that each region can work in order to achieve its goal. The diffusion
mechanism only relies on the variation of X.
Another hypothesis is to consider that a small fringe close to the borderline (area
in dotted line, Fig. 3) takes part in the diffusion of wealth, by the leveling-out of
standards of living of the two neighboring border fringes (Fig. 3). Since we do not
have any information on the spatial distribution of the populations inside a region,
we must put forward the hypothesis of a uniform distribution. Consequently, we use
a simple proportionality parameter, called the diffusion rate, the value of which can
be set within the user interface. This parameter rate k (of surface area, population,
and wealth) is all at once, since we consider them as uniformly distributed over the
region’s surface area.
Figure 3. Practical implementation of growth-diffusion rate
In order to model the diffusion between two regions i and j, we then introduce
the coefficient kij which is the surface area’s proportion i matching the intersection
between the border fringe defined by k and the proportion pij of its borderline land
shared by the region j , defined by
, where lij is the borderline’s
length between i and j.
We then have :
ijij pkk .
If the wealth on the two sides of the border fringe between i and j was evenly
distributed like connected areas, we would obtain a leveled-out standard of living
(which is not the average of the two previous standards), defined by:
We can then define the variation dXij (positive if it emits or negative if it
receives) of the diffusion from the region i towards the region j during a short lapse
of time dt as being proportional to the concerned population (kijPi) and proportional
to the difference between the current standard of living (Yi) and the (local) aim of
leveling-out (Yij) of standards of living i and j. This can be translated into the
The value of K is set internally (since we can already play on k).
By adding the border fringes of the region i, we note:
One should notice that this diffusion is, by construction, preservative of the
(because one can verify easily that for any couple (i, j) we have:
dXij + dXji = 0)
Moreover, the variable Xi is subjected to an a priori exponential internal growth,
Internal growth is adjustable, either individually region by region through the
attribute table, either on the whole as being the same for all regions with the help of
a setting determined by the user within.
The final growth-diffusion equation is thus given by:
dtYYPkKtXCtXdttX ijiiijiiii ))(.)(.()()(
The lapse of time for the discretization of growth and diffusion processes are
small compared to redistributing flows, because they correspond to continuous
processes. We have selected the month as lapse of time, that also matches the time
unit that we chose, so dt =1. (Ci is then the twelfth of the annual growth rate).
The equation with this lapse of time is then written:
)(.)()1()( ijiiijiii YYPkKtXCdttX
4. Europe 2025 : Which scenario from which policy?
To assess the weight of political cohesion in regional trajectories, simulations
were performed with the GeoCells platform. These simulations were based on the
one hand on the settings of allocations Funds and, on the other hand on
neighborhood effects. The two scenarios presented below ask questions about the
effectiveness of the cohesion policy and the dilemma between competitiveness and
equity. This dilemma can be read in the Treaty of Rome (1957) and the Single
European Act (1987) where it says that the EU has to support the growth and the
job creation in Member states and least developed regions.
The first scenario (simu 1) is the one of free competition between regions
without the intervention of Cohesion Policy (table 1). It is tantamount to abolishing
European “interventionism” and to “renationalizing” aid, just as recommended in
the Sapir Report. “An Agenda for a Growing Europe”, also called The Sapir Report,
is a report on the economy of the European Union edited by a panel of experts under
the direction of André Sapir and published in July 2003. The report follows an
initiative by Romano Prodi, President of the European Commission, notably to
analyze the Lisbon Strategy. According to the experts of this report, Cohesion
Policy and other community interventions do not contribute in an easily measurable
way to the convergence of the regions. The results obtained by the countries of the
EU remain dependent on their good governance, which leads the experts of this
report to write the following recommendation: “there is a solid argument for the
new EU convergence policy to focus on countries, rather than on regions”.
Considering the European budgetary constraints, the report recommends an
important reduction of Funds intended for the Cohesion Policy. The simulations
include a low percentage of Cohesion Policy in the EU budget. Almost all regions
can apply for the Cohesion Policy because the threshold of allocation of Cohesion
Policy of 90 % of the average GDP per capita of the EU. We observe that
disadvantaged regions catch up slowly and the sigma convergence indicates
. The distribution of wealth is more non-egalitarian than the scenario of
The second scenario (Simu 2) has as its goal territorial equity (Table 1).
Territorial equity includes ideas of parity of treatment, equality of access, and, more
generally, solidarity between regional organizations in terms of public action,
especially by implementing corrective measures as far as resources and facilities are
concerned. The scenario consists of endowing each region with a measure of
autonomy and the necessary conditions for development. Cohesion Policy are used
alone, by increasing the percentage devoted to regional policy to 35% of EU budget,
and by retaining the actual threshold of allocation of Cohesion Policy to 75% of the
average GDP per capita of the EU.
Table 1. Indicators in Scenarios of Cohesion Policy in Europe in 2025
Simu 1 :
Simu 2 :
The sigma convergence refers to a reduction in the dispersion of levels of income across
economies. Here there is an increase of disparities among regions because of positive result
• Cohesion Policy – per-
centage of the EU
• Threshold of alloca-
tion of Cohesion Policy
(GNI/gross national in-
come as percent of EU
• Beta convergence
• Sigma convergence
• GDP diffusion rate
• Gini index
• Moran index
• Priorities for the cohe-
Integration of the less economically de-
• Public policies
Renationalization of aid
Increase Cohesion Policy total budget
• Mechanism for pro-
Liberalization and competition in-
The measure of convergence based on the evolution of the standard deviation
(sigma convergence) gives the most valuable result for the scenario of equity based
upon increasing the budget for regional policy, and the prospect of catching-up (beta
convergence) is more credible with the scenario of equity as well. With this policy
orientation, every region of each country reacts positively to territorial solidarity
programs. In accordance with the results in terms of beta and sigma convergence,
simulation 2 brings out a better result in terms of Territorial Cohesion, mitigating
significatively regional disparities across EU.
The cartography of these scenarios gives concrete expression to the impact on
geographic distribution of growth chosen by each parameter setting (Fig. 4). We
have measured local concentrations through the Getis-Ord statistics. This index
allows the identification of spatial clusters (or "local pockets"). A positive value
will indicate a spatial concentration of GDP per capita (called a "hot spot"), while a
negative value of that index is associated with spatial concentration of low value of
GDP per capita (a “cold spot”). Two main patterns of clusters can be shown. Within
the competitive scenario, the “Pentagon” (cluster of prosperous regions) and regions
bordering this cluster are strongly linked to each other; unfortunately many regions
of formerly socialist countries remain far behind. This scenario produces the
phenomena of the clustering of prosperous regions very often from metropolitan
regions (South of England, Parisian Basin, North West of Italy) whereas poor
regions do not manage to progress of their backwardness. Representative of a non-
egalitarian growth, this phenomena reveals a certain effectiveness at national level,
but establishes itself as less homogeneous at European level. Growth takes place by
clusters of regions, but the development gaps are not on the whole being closed (low
beta convergence). On the other hand, the scenario of equity highlights the progress
of the convergence of GDP. It allows CEECs regions to catch up while allowing the
Pentagon to continue to grow. This hypothesis gives a negative sigma convergence
with a low dispersion of incomes between regions, since poorer regions saw their
GDP per capita rise, but, not at the same rhythm. CEECs regions located closest to
the former Iron Curtain seem to be progressing faster.
Fig. 4. Cartography of Scenario 1 and 2 - Spatial statistics
In addition to the two indicators of convergence (beta and sigma) – in theory
complementary and often referenced in the literature in regard to regions’
convergence – the introduction of a contiguity-based growth propagation variable
changed the expected scenarios which stood as a basis for EU policies. This
introduction of spatial interaction by neighborhood transforms the deterministic
projections of the EU policies into a system of regional units reacting according to
a multi-scalar complexity. The process accounting neighborhood effects reveals the
potential for a spatial diffusion process to occur under the assumptions given in each
The budget of the European regional policy has always been the second largest
item of expenditure in the EU, far behind that of the Common Agricultural Policy.
With the new programming period 2007-2013, the budget was brought to the
forefront because of the efforts related to the 2004 and 2007 enlargements.
Achieving competitiveness of the regions included in the Lisbon strategy requires
building development strategies that enhance regional strengths and overcome
weaknesses and regional gaps. To meet the challenges of globalization, the EU has
included the concept of competition in the 2007-2013 programming period for the
Cohesion Policy. Meanwhile, the EU continues to pursue its objective of solidarity
between regions and countries. This dilemma can be answered by the concept of
. This concept refers to a development of a polycentric and
balanced urban system, and strengthening of the partnership between urban and
rural areas, so as to create a new urban-rural relationship. It includes the promotion
of integrated transport and communication, which support the polycentric
development of the EU territory, so that there is gradual progress towards parity of
access to infrastructure and knowledge. Implicitly, this principle implies the
presence of "centers" that spread their prosperity to their neighborhoods (hence the
need to introduce neighborhood effects in GeoCells) while continuing to help the
less economically developed regions to be competitive vis-à-vis the wealthier
would combine greater European competitiveness with an increase in prosperity of
peripheral regions in order to catch up. The spatial dimension of European public
action is an opportunity to resolve these contradictions. The territorialization of
public policies for regional development (which consists of differentiating policy
applications for different regions) coupled with a polycentric planning can allow a
difficult compromise between equity and competitiveness.
The objective of this chapter was not to provide an answer on how the Cohesion
Policy should be used (axiological neutrality) but to clarify issues for the future of
European cohesion policy. This clarification is necessary to understand the
geographic organization of economic inequality and regional development. The two
scenarios that have been demonstrated in this study show that the political choices
between equity and competitiveness have a profound impact on territorial
development. These choices in structural funding investment produce very different
economic and spatial configurations. Not only the political orientation can influence
outcome, but other factors can have a significant impact on territorial cohesion. Both
pre-determined (i.e. programming policies, historical factors) and random
(neighborhood effects, diffusion of regional growth) factors affect the dynamics of
regional growth and convergence. Because each region has a unique trajectory
“Promote a harmonious and well-balanced development of the EU’s territory”, European
Sapir Report advocates this but stopping aid to regions in an earlier stage of economic
development, thus not allowing these regions to be competitive vis-a-vis the wealthiest
based not only upon Cohesion Policy but also upon random factors, it is impossible
to directly link Cohesion Policy alone to regional economic growth.
At this stage of our research, it would be helpful to use an input-output model
as an extension for future work. The input-output model would represent the
sectoral diffusion of the funding (underlying processes) and the simulation could
represent the resulting geographic diffusion/interactions. The goal would be to
explore the logical consequences of assumptions based on neighborhood effects, to
complete them with the simulation results so get to know the reality and act more
effectively on it.
Aghion, P., & Cohen, E.. (2004). Education et Croissance. Paris : La Documentation Française.
Baumont, C. (1998). Economie géographique et intégration régionale : Quels enseignements pour
les Pays d’Europe Centrale et Orientale ? LATEC – Economics Working Paper (1991-2003), 1998-
Baumont, C., Ertur, C., & Le Gallo, J. (2002). The European Regional Convergence Process, 1980-
1995: Do Spatial Regimes and Spatial Dependence Matter? Economics Working Paper Archive -
Bergs, R. (2001). EU Regional and Cohesion Policy and Economic Integration of the Accession
Countries. Discussion Paper, Policy Research & Consultancy.
Bourdin, S. (2013). Une mesure spatiale locale de la sigma-convergence pour évaluer les disparités
régionales dans l'Union européenne, Région et Développement, 37.
Bradley, J. et Untiedt, G. (2007). Do economic models tell us anything useful about Cohesion
Policy impacts? A comparison of HERMIN, QUEST and ECOMOD. Working Paper, 3, GEFRA
- Gesellschaft fuer Finanz- und Regionalanalysen.
Bradley, J., Morgenroth, E. et Untiedt, G. (2003). Macro-regional evaluation of the Structural
Funds using the HERMIN modelling framework. ERSA Congress University of Jyväskylä - Fin-
Bradley, J. (2002). An Examination of the Ex-post Macroeconomic Impacts of CSF 1994-99 on
Objective 1 Countries and Regions. Greece, Ireland, Portugal, Spain, East Germany and Northern
Ireland. Final Report, ESRI & GEFRA, Dublin.
Bradley, J., Herce, J-A. et Leonor, M. (1995). Modelling in the EU periphery : The HERMIN
project. Economic Modelling, 12(3), 219-22.
Cappelen, A., Castellacci, F., Fagerberg, J. et Verspagen, B. (2003). The Impact of EU Regional
Support on Growth and Convergence in the European Union. Journal of common market studies,
Coudroy de Lille, L. (2009). Les nouveaux territoires polonais. In Jean, Y., & Baudelle, G. (Ed.),
L’Europe, aménager les territoires, Paris : Armand Collin.
Cour, P. et Nayman, L. (1999). Fonds Structurels et disparités régionales en Europe. La lettre du
Dall'erba, S., & Le Gallo, J.. (2008). Regional convergence and the impact of European Cohesion
Policy over 1989-1999: a spatial econometric analysis. Papers in Regional Science, 87, 219-244.
Dall'erba, S. and Hewings, G.J.D. (2009). European regional development policies: The trade-off
between efficiency-equity revisited. Connections, 5, 73–84.
Dall'erba, S., Guillain, R., and Le Gallo, J. (2009). Impact of Cohesion Policy on regional growth:
How to reconsider a 9-year-old black box. Région et Développement, 30, 77-100.
Dühr, S., Nadin, V., Colomb, C. (2009). European spatial planning: territorial development, co-
operation and EU spatial policy. Londres : Routledge.
Ederveen, S., Groot, H.L.F. et Nahuis, R. (2006). Fertile Soil for Structural Funds? A Panel Data
Analysis of the Conditional Effectiveness of European Cohesion Policy. Kyklos, 59(1), 17-42.
Ederveen, S. et Gorter, J. (2002). Does European Cohesion Policy Reduce Regional Disparities?
An Empirical Analysis. CPB Working Paper, 15, CPB Netherlands Bureau for Economic Policy
Elissalde, B., Goyat, D., & Langlois, P. (2009). GeoCells model : European Cohesion Policy and
regional interactions, which convergences for the European regions? Revue électronique Cyber-
geo. doi : 10.4000/cybergeo.22388.
Ertur, C., and Le Gallo, J. (2008). Regional growth and convergence : heterogenous reaction versus
interaction in spatial econometric approaches. oai : hal.archives-ouvertes.fr:hal-00463274.
ESPON Project 2.2.1. (2005). The territorial effects of the Cohesion Policy, Final report. Esch-
sur-Alzette : Espon Coordination Unit and Nordregio.
European Commission (1998). Resolution on regional planning and the European Spatial Devel-
opment Perspective, Brussels.
European Commission. (2006). Cohesion policy and cities : the urban contribution to growth and
jobs in the regions, Brussels.
European Commission. (2008a). Growing regions, growing Europe, 5th cohesion report. Brussels.
European Commission. (2008b). Green Paper on territorial cohesion. Brussels : DG Regio.
European Parliament. (2007). Regional disparities and cohesion : what strategies for the future ?
European Parliament. (2008). Shrinking regions : a paradigm shift in demography and territorial
Ezcurra, R. & Rapún, M. (2006). Regional disparities and national development revisited, the case
of western Europe. European Urban and Regional Studies, 13 (4), 355-369.
Fayolle, J., & Lecuyer, A. (2000). Croissance régionale, appartenance nationale et fonds structurels
européens. Un bilan d’étape. Revue de l’OFCE, 73, 161-196.
Getis, A. (1991). Spatial interaction and spatial autocorrelation: a cross-product approach. Envi-
ronment and Planning, 23, 1269-1277.
Gorzelak, G., Maier, G., & Petrakos G. (2010). Integration and transition in Europe: the economic
geography of interaction. London : Routlege.
Islam, N. (2003). What have we learnt from the convergence debate ? Journal of Economic Sur-
veys, 17, 309-362.
Jouen, M. (2008). La cohésion territoriale, de la théorie à la pratique. Notre Europe.
Kelber A. (2010). La politique de cohésion et les nouveaux États membres de l’Union européenne.
Bulletin de la Banque de France, 181 (3), 1-13.
Kilper, H. (Ed.). (2009). New disparities in spatial development in Europe. German Annual of
Spatial Research and Policy, Springer.
Lackenbauer, J. (2006). Equity, efficiency, and perspectives for cohesion policy in the enlarged
European Union. BERG Velag, Bamberg.
Le Gallo, J. (2004). Space-time analysis of GDP disparities among European regions : a Markov
chains approach. International Regional Science Review, 27, 138-163.
Magnier, P. (2004). Fonds structurels européens et politiques régionales. Commissariat général
du plan, La Documentation française.
PANORAMA (2010). Evaluating Regional Policy : Insights and results. Inforegio, 33.
Rey, S.J., & Janikas, M.V. (2005). Regional convergence, inequality and space. Journal of Eco-
nomic Geography, 5 (2), 155-176.
Roeger, W. et in 't Veld, J. (1997). QUEST II. A Multi-Country Business Cycle and Growth Model.
European Economy - Economic Papers, 123, Directorate General Economic and Monetary Af-
fairs, European Commission.
Roeger, W. et in 't Veld, J. (2004). Some selected simulation experiments with the European com-
mission's QUEST model. Economic Modelling, 21(5), 785-832.
Sapir J., Aghion P., Bertola G., Hellwig M., Pisani-Ferry J., Rosati D., Viñals J., Wallace H., Buti
M., Nava M. et Smith P.M. (2003). An Agenda for a Growing Europe: The Sapir Report, ed.
Oxford University Press, Oxford.
Stryjakiewicz T. (2007). La nouvelle géographie de la Pologne dans le contexte de la transition
économique et sociale et de l’élargissement de l’Union européenne. L’Information Géographique,
Varga, J. et in 't Veld, J. (2011). A model-based analysis of the impact of Cohesion Policy ex-
penditure 2000-06: Simulations with the QUEST III endogenous R&D model. Economic Model-
ling, 28(1-2), 647-663.
Williamson, J. G. (1965). Regional Inequality and the Process of National Development : A De-
scription of the Patterns, Economic Development and Cultural Change, 13, 1-84.