You are on page 1of 34

Economic Papers are written by the Staff of the Directorate-General for Economic and Financial

Affairs, or by experts working in association with them. The Papers are intended to increase awareness
of the technical work being done by staff and to seek comments and suggestions for further analysis.
The views expressed are the author’s alone and do not necessarily correspond to those of the European
Commission. Comments and enquiries should be addressed to:

European Commission
Directorate-General for Economic and Financial Affairs
Publications
B-1049 Brussels
Belgium
E-mail: Ecfin-Info@ec.europa.eu
























This paper exists in English only and can be downloaded from the website
http://ec.europa.eu/economy_finance/publications

A great deal of additional information is available on the Internet. It can be accessed through the
Europa server (http://europa.eu )




KC-AI-09-371-EN-N
ISBN 978-92-79-11182-2
ISSN 1725-3187
DOI 10.2765/2997



© European Communities, 2009
A Model-based Assessment of the Macroeconomic
Impact of EU Structural Funds on the New Member
States
Janos Varga and Jan in ’t Veld

March 9, 2009
Abstract
This paper gives a model-based analysis of the potential macro-economic
impact of European Union Structural and Cohesion Funds payments on the
economies of the new member states. The model used is a four-region DSGE
model with human capital accumulation and endogenous technological change.
The framework that we adopt is the Jones (2005) extension of the endogenous
growth model, which uses a variety approach for modelling knowledge invest-
ment. The EU funds average around 1.5 percent of GDP and are used for
investment in infrastructure, human capital and R&D. The model simulations
show this can lead to signi…cant gains in output, both in the short as well as
in the long run.
JEL Classi…cation: E62, H50, O11, O41
Keywords: Fiscal transfers, Structural Funds, public investment, DSGE
modelling.

We like to thank Dalia Grigonyte for providing us with the dataset. The views expressed in this paper
are those of the authors and should not be attributed to the European Commission. Correspondence:
Janos Varga, European Commission, DG ECFIN Economic and Financial A¤airs, B-1049 Brussels, Bel-
gium, email: janos.varga@ec.europa.eu ; Jan in ’t Veld, European Commission, DG ECFIN Economic
and Financial A¤airs, B-1049 Brussels, Belgium, email: jan.intveld@ec.europa.eu
1 Introduction
The European Union’s Cohesion Policy provides a framework for large …scal transfers from
the richer EU Member States to the countries and regions that lag behind in terms of
income per capita. The Cohesion and Structural Funds are, after agricultural spending,
the second largest item in the current EU budget, accounting for 34 per cent of total
appropriations for commitments. The resources are targeted on investment in physical
and human capital, and designed to increase economic and social cohesion among member
states, enhancing a faster catch-up process of the less developed member states. For
eligibility for Objective One status the region’s GDP per capita must be less than 75 per
cent of the EU average, although special ad hoc arrangements exist.
At the time of the creation of the Single Market programme, it was argued that poorer
member states had to be compensated for exposing their weaker economies to more com-
petition in the internal market, and with EU enlargement in 2004 there has been a sharp
increase in EU Structural Funds. The funds are also seen as an expression of solidar-
ity between member states and a commitment to redistribution of incomes considered
necessary for an e¤ective functioning of the Union’s political institutions (Gros (2008)).
Economic theory predicts unambiguous bene…ts from investment in infrastructure and
human capital and there is empirical evidence supporting this. However, ex-post evalu-
ation studies of large EU transfers in the past generally give only mixed support for EU
Cohesion Policy. Many of the assisted regions have remained relatively poor and growth
regressions augmented with Structural Fund variables show no signi…cant impact from
these transfers (see Ederveen et al. (2002) and Ederveen et al. (2006)). Model based stud-
ies, frequently used for the purposes of ex-ante evaluation, usually show larger impacts
of the demand stimulus as well as positive supply-side e¤ects on GDP in the long run.
However, these studies have often been based on macroeconometric models in which the
demand side is essentially keynesian in nature and no-crowding-out appears (e.g. Bradley
et al. (2007)). On the whole, empirical evidence on the positive impact of Structural
1
Funds on convergence of poorer regions remains inconclusive.
It should be recognised that the objective of cohesion policy is much wider than eco-
nomic growth per se. A cursory look at the …elds of interventions of CSF shows a wide
range of policies promoting environmental protection and risk prevention, tourism, cul-
ture, urban and rural regeneration, improving social inclusion of less-favoured persons.
While these policies undoubtedly promote social cohesion, they are not directly enhancing
the growth potential of an economy, at least not in a unequivocal and directly measurable
way. Nevertheless, economic convergence is the ultimate objective of cohesion policy and
the eligibility criterium for Objective One status is clearly de…ned in terms of income per
capita. Hence, it is important to examine whether Structural Funds can promote growth
and enhance a faster catching up of the less developed member states.
This paper uses the QUEST III model to evaluate the potential impact of Structural
and Cohesion Fund programmes for the new member states for the period 2007-2013.
While it should be clear that an ex-ante model-based evaluation cannot provide a con-
clusive answer to the question whether structural funds will accelerate the economic con-
vergence of poorer regions, an analysis based on a model that incorporates the channels
through which these policies can promote growth can provide an estimate of the poten-
tial impact, assuming an optimal use of the funds.
1
The model we use in this study
is based on a four region dynamic general equilibrium model with human capital ac-
cumulation and endogenous technological change. It has been used extensively for the
analysis of structural reforms in the EU (the Lisbon Strategy for Growth and Jobs) (see
Roeger et al. (2008)) and is particularly suitable for an evaluation of the type of struc-
tural policies that form the core of Structural Funds interventions. The model belongs
to the new class of micro-founded DSGE models that are now widely used in economic
policy institutions. The model incorporates productive public investment that captures
the productivity-enhancing e¤ects of infrastructure investment. The model employs the
product variety framework proposed by Dixit and Stiglitz (1977) and applies the Jones
1
In a comprehensive study of absorption problems, Herve and Holzmann (1998) list many factors that
could lead to sub-optimal use of …scal transfers and argue that the scope for problems related to e.g.
rent-seeking may be very high.
2
(1995) semi-endogenous growth framework to explicitly model the underlying develop-
ment of R&D. The endogenous modelling of R&D allows us to analyse the impact of
R&D promoting policies on growth and the endogeneity of human capital accumulation
can capture the e¤ects of policies promoting vocational education and training.
The paper is organized as follows. The next section brie‡y discusses the Structural and
Cohesion Funds programmes and the …scal transfers involved. Section 3 describes those
features of the model which enable us to carry out the impact assessment of the …scal
transfers. The model results depend crucially on 1. the assumptions on the productive
impact of additional public capital and 2. on how the skill e¢ciencies are a¤ected by
human capital investment. The simulation results and a sensitivity analysis with respect
to these assumptions are discussed in Section 4. Finally, Section 5 concludes.
2 The Structural and Cohesion Funds programme
2007-2013
For the period 2007 to 2013, Structural and Cohesion Funds programmes for the New
Member States amount to a total budget of 173.9 billion euros (in 2008 prices). But
because past experience in previous programme periods have shown considerable delays
in payments, typically continuing for up to two more years, the proposed payments are
spread over a 9 year period lasting till 2015. On the basis of payment pro…les in pro-
gramming prices, assuming an in‡ation correction of 2 per cent per year, and using the
Commission’s nominal GDP projections for 2008-15, we calculate in Table 1 the proposed
annual payment pro…le in terms of GDP for 2007-2015.
The …elds of interventions are divided into …ve main categories (plus an additional
technical assistance category). Figure 1 shows the distribution of the total budget for
each of the New Member States.
As the …gure shows, infrastructure investment receives the largest share of funds, more
than 60% of the total budget for most NMS, while investments in R&D and human capital
are the second or third largest categories. The …elds of intervention cover a wide range of
3
Table 1: Payment Pro…le 2007-2015
Country 2007 2008 2009 2010 2011 2012 2013 2014 2015
Bulgaria 1.1 1.1 1.1 1.4 1.7 2.0 2.2 2.6 1.9
Czech Republic 1.2 0.9 0.8 1.3 1.7 1.8 1.9 2.4 1.7
Estonia 1.1 1.2 1.3 1.7 2.0 1.9 1.9 1.9 1.6
Cyprus 0.2 0.2 0.2 0.3 0.3 0.4 0.4 0.7 0.4
Latvia 1.0 1.0 1.0 1.0 0.9 1.7 2.4 2.0 2.0
Lithuania 1.1 1.3 1.5 1.4 1.4 1.8 2.2 2.7 1.9
Hungary 1.0 1.3 1.6 2.0 2.4 2.7 2.9 3.0 2.6
Malta 0.5 0.5 0.4 0.8 1.2 1.9 2.6 2.2 2.4
Poland 1.1 0.9 0.7 1.1 1.5 1.8 2.2 2.7 1.9
Romania 0.7 0.8 0.8 1.0 1.2 1.4 1.5 1.7 1.3
Slovenia 0.6 0.8 1.0 1.2 1.3 1.1 0.9 1.3 0.8
Slovakia 1.1 1.1 1.1 1.3 1.4 1.7 1.9 2.3 1.6
All NMS 1.0 0.9 0.9 1.3 1.5 1.8 2.0 2.4 1.8
Source: own estimates. The payment pro…le in programming prices was provided
by DG REGIO with an assumption of 2% in‡ation. GDP projections are obtained
from the AMECO database and DG ECFIN’s potential growth calculations.
policy programmes, details of which are shown in the annex (Table A1).
3 Modelling the impact of Structural Funds
The model used in this analysis is an extended version of the QUEST III R&D model.
The model consists of four regions: the block of new member states, the Euro area and
non-Euro area old member states and the rest of the world. This section describes the
main channels through which the model captures the impact of the di¤erent interventions
on growth. A more detailed description of the model can be found in Roeger et al. (2008).
3.1 The Model
The model economy is populated by households, …nal and intermediate goods producing
…rms, a research industry, a monetary and a …scal authority . In the …nal goods sector
…rms produce di¤erentiated goods which are imperfect substitutes for goods produced
abroad. Final good producers use a composite of domestic and imported intermediate
4
Figure 1: Shares of payments by …elds of interventions
0%
20%
40%
60%
80%
100%
BUL CZR EST LAT LIT HUN POL ROM SLO SLK NMS
0%
20%
40%
60%
80%
100%
Infrastructure RTD Services Industry Tech. Assis Human Capital
goods and three types of labour - (low-, medium- and high-skilled). Households buy the
patents of designs produced by the R&D sector and license them to the intermediate goods
producing …rms. The intermediate sector is composed of monopolistically competitive
…rms which produce intermediate products from rented capital input using the designs
licensed from the household sector. The production of new designs takes place in research
labs, employing high skilled labour and making use of the existing stock of domestic
and foreign ideas. Technological change is modelled as increasing product variety in the
tradition of Dixit and Stiglitz (1977).
The model distinguises two types of households. The …rst group of households have
access to …nancial markets where they can buy and sell domestic and foreign assets (gov-
ernment bonds), accumulate physical capital which they rent out to the intermediate
sector, and they also buy the patents of designs produced by the R&D sector and license
them to the intermediate goods producing …rms. These household members o¤er medium-
and high-skilled labour services. Another share of households is liquidity-constrained.
These households cannot trade in …nancial and physical assets and consume their dispos-
5
able income each period. Members of liquidity constrained households o¤er low-skilled
labour services only. For each skill group we assume that both types of households sup-
ply di¤erentiated labour services to unions which act as wage setters in monopolistically
competitive labour markets. The unions pool wage income and distribute it in equal
proportions among their members. Nominal rigidity in wage setting is introduced by
assuming that households face adjustment costs for changing wages.
Below we describe in some more detail the supply side of the model and the …scal side,
which constitute the key elements for modelling the Structural Funds interventions. One
extension to the model made here is an explicit formulation of human capital accumulation
following Jones (2002) in order to account for the signi…cant part of Structural Fund
investments in various human resource programmes. In calibrating the model, we follow
the literature of dynamic general equilibrium modelling and set the key steady-state ratios
equal to their empirical counterparts for each region. While the calibration of the main
steady state ratios (private consumption to output, investment to output, etc.) is based
on EUROSTAT and OECD data, the remaining structural parameters and variables are
adopted from the available estimates in empirical studies (see Ratto et al. (2008)) or tied
down by the equations of the model.
2
3.1.1 Final goods production and public capital
We account for the productivity-enhancing e¤ect of infrastructure investment via the
following aggregate …nal goods production function:
1
t
= ¹
(1c)
(
1
0
1
)
t
_
1
1
t
_
1c
_
1
Y,t
_
c
_
1
G
t
_
c
C
1C
Y
. where
¹
I

i=1
r
i,t
= 1
1
t
(1)
The …nal good sector uses a labour aggregate (1
Y,t
) and intermediate goods (r
i,t
)
using a Cobb-Douglas technology, subject to a …xed cost 1C
Y
. Our formulation assumes
2
See the appendix for the description and the calibration of knowledge production and human capital
accumulation equations.
6
that investment in public capital stock (1
G
t
) increases total factor productivity with an
exponent of c
G
. Final output ( 1
t
) is produced using ¹
t
varieties of intermediate inputs
with an elasticity of substitution 1,(1 o). One unit of intermediate goods is produced
from one unit of private capital (1
1
t
), therefore in a symmetric market framework the
total output of the intermediate sector amounts to the total private capital stock as
¹
I

i=1
r
i,t
= ¹
t
r
t
= 1
1
t
.
Public infrastructure investment (1
G
t
) accumulates into the public capital stock 1
G
according to
1
G
t
= (1 o
G
)1
G
t1
+1
G
t
(2)
where o
G
, the depreciation rate of public capital is set at 4 per cent.
Infrastructure investment is assumed to be proportional to output
1
G
t
= (1Go
t
+
1G
t
)1
t
(3)
where
1G
t
is an exogenous shock to the share of government investment (1Go
t
). It is
through this shock that we simulate the increase in infrastructure investment.
3.1.2 Intermediate production and the R&D sector
The intermediate sector consists of monopolistically competitive …rms which have entered
the market by buying licenses for design from domestic households and by making an
initial payment 1C
¹
to overcome administrative entry barriers. Capital inputs are also
rented from the household sector for a rental rate of i
1
t
. Firms which have acquired a
design can transform each unit of capital into a single unit of an intermediate input.
Intermediate goods producing …rms sell their products to domestic …nal good producers.
In symmetric equilibrium the inverse demand function of domestic …nal good producers
7
is given as
jr
i,t
= j
t
(1 c)1
_
¹
I

i=1
_
r
)
i,t
_
0
_
1
_
r
i,t
_
01
(4a)
where j
t
is the inverse gross mark-up of the …nal goods sector.
Each domestic intermediate …rm solves the following pro…t-maximisation problem.
11
a
i,t
= max
a
.,I
_
jr
i,t
r
i,t
i
1
t
1
C
t
/
i,t
i
¹
1
¹
t
1C
¹
_
. (5)
subject to a linear technology which allows to transform one unit of e¤ective capital
(/
i
nccj) into one unit of an intermediate good
r
i
= /
i
. (6)
The no-arbitrage condition requires that entry into the intermediate goods producing
sector takes place until
11
a
i,t
= 11
a
t
= i
¹
t
1
¹
t
+:
t
1C
¹
t
(7)
or equivalently, the present discounted value of pro…ts is equated to the …xed entry costs
plus the net value of patents
1
¹
t
1
1 t
1
t
_
1 o
¹
_
+t
¹
+1C
¹
=
1

t=0
t

)=0
_
1
1 +:
t+)
_
11
a
t+t
. (8a)
For an intermediate producer, entry costs consist of the licensing fee i
¹
t
1
¹
t
for the design
or patent, which is a prerequisite of production of innovative intermediate goods, and the
…xed entry cost 1C
¹
.
Innovation corresponds to the discovery of a new variety of producer durables that
provides an alternative way of producing the …nal good. The R&D sector hires high-skilled
labour 1
¹,t
and generates new designs according to the following knowledge production
function:
8
¹
t
= i¹
v
t1
¹
ç
t1
1
A
¹,t
. (9)
In this framework we allow for international R&D spillovers following Bottazzi and
Peri (2007). Parameters ¬ and c measure the foreign and domestic spillover e¤ects from
the aggregate international and domestic stock of knowledge (¹

and ¹) respectively.
Negative value for these parameters can be interpreted as the "…shing out" e¤ect, i.e.
when innovation decreases with the level of knowledge, while positive values refer to the
"standing on shoulders" e¤ect and imply positive research spillovers. Note that c = 1
would give back the strong scale e¤ect feature of fully endogenous growth models with
respect to the domestic level of knowledge. Parameter i can be interpreted as total factor
e¢ciency of R&D production, while ` measures the elasticity of R&D production on the
number of researchers (1
¹
). The international stock of knowledge grows exogenously at
rate q
¹
u. We assume that the R&D sector is operated by a research institute which
employs high skilled labour at their market wage \
1
. We also assume that the research
institute faces an adjustment cost of hiring new employees and maximizes the following
discounted pro…t-stream:
max
1
/,I
1

t=0
d
t
_
1
¹
t
¹
t
\
1
t
1
¹,t

¸
¹
2
\
1
t
1
2
¹,t
_
(10)
Therefore the …rst order condition implies:
`1
¹
t
¹
t
1
¹,t
= \
1
t

¹
_
\
1
t
1
¹,t
d
t
\
1
t+1
1
¹,t+1
_
(11)
where d
t
is the discount factor.
3.1.3 Human capital accumulation
The labour aggregate 1
Y,t
is composed of three skill-types of labour force:
9
1
Y,t
=
_
:
1
¬
1
1
_
/
1
t
1
1
t
_
¬
1
1
¬
1
+:
1
¬
1
A
_
/
A
t
1
A
t
_
¬
1
1
¬
1
+:
1
¬
1
1,Y
_
/
1
t
1
1Y
t
_
¬
1
1
¬
1
_
¬
1
¬
1
1
. (12)
Parameter :
c
is the population share of the labour-force in subgroup : (low-, medium-
and high-skilled), 1
c
denotes the employment rate of population :, /
c
t
is the corresponding
accumulated human capital (e¢ciency unit), and o
1
is the elasticity of substitution be-
tween di¤erent labour types
3
. An individual’s human capital is produced by participating
in education and
c
t
represents the amount of time an individual spends accumulating
human capital :
/
c
t
= /
c
c
ç
s
I
. · 0 (13)
The exponential formulation used here adapts Jones (2002) into a disaggregated skill-
structure by incorporating human capital in a way that is consistent with the substantial
growth accounting literature with adjustments for education
4
. The · parameter has been
studied in a wealth of microeconomic research. Interpreting
c
t
as years of schooling,
the parameter corresponds to the return to schooling estimated by Mincer (1974). The
labour-market literature suggests that a reasonable value for · is 0.07, which we apply
here. Investments in human capital can then be modelled by increasing the years of
schooling (
c
t
) for the respective skill-groups.
3.2 The government budget constraint
For the government sector various expenditure and revenue categories are separately mod-
elled. On the expenditure side we assume that government consumption (G
t
), government
transfers (11
t
) and government investment (1
G
t
) are proportional to GDP and unemploy-
ment bene…ts (11`
t
) are indexed to wages. The government provides subsidies (o
t
)
3
Note that high-skilled labour in the …nal goods sector L
HY
t
is total high-skilled employment minus
the high-skilled labour working in the R&D sector ( L
A;t
).
4
See Barro and Sala-I-Martin (1995).
10
on physical capital and R&D investments in the form of a tax-credit and depreciation
allowances.
Government revenues (1
G
t
) are made up of taxes on consumption as well as capital
and labour income. Fiscal transfers for NMS received from the EU are denoted by CCH
t
(which is negative for the net contributors). There is a lump-sum tax (1
1S
t
) used for
controlling the debt to GDP ratio according to the following rule
1
1S
t
= t
1
_
1
t1
1
t1
/
T
_
+t
111

_
1
t
1
t
_
(14)
where /
T
is the government debt target, t
1
and t
111
are coe¢cients. Therefore,
government debt (1
t
) evolves according to
1
t
= (1 +:
t
) 1
t1
+1
C
t
G
t
+11
t
+11`
t
+o
t
1
G
t
1
1S
t
CCH
t
(15)
It is assumed that the additional contributions to the EU budget are …nanced in the
donor countries through an increase in lump-sum taxes.
Cohesion policy programmes are subject to the condition of additionality and co-
…nancing. Additionality requires that Structural Funds are additional to domestically-
…nanced expenditure and are not used as a substitute for it. The co-…nancing principle
means the EU provides only matching funds to individual projects that are part of the
operational programmes and that the EU funds are matched to a certain extent by domes-
tic expenditure. The problem with de…ning a proper benchmark means that in practice
this principle of additionality is hard to verify and is thus not always binding. Member
States are not required to create new budgetary expenditure to co-…nance cohesion policy
support. Existing national resources that were used to …nance similar areas of interven-
tions (and are thus concerned by the additionality requirement) can be ’earmarked’ to
co-…nance Structural Fund transfers. Total spending increases only by the amount of
Structural Fund transfers.
More formally, assume a co…nancing rate of c, i.e. the EU transfer CCH
t
has to
be matched by domestically-…nanced expenditure c.CCH . The additionality and co-
11
…nancing principles can be expressed as the following condition for total government
spending in a bene…ciary country:
1C11A1
t
= CCH
t
+ max(1A1
0
. c CCH
t
) (16)
where 1C11A1
t
is total expenditure, CCH
t
is the …scal transfer received from the EU
cohesion funds, 1A1
0
domestically–…nanced expenditure in the counterfactual situation
(without Structural and Cohesion Funds), and c is the co-…nancing rate. Examining the
additionality tables of Member States, it is apparent that national public expenditure
concerned by additionality usually exceeds the co-…nancing needs by far. In this case
1A1
0
c CCH
t
, and total expenditure is given by
5
1C11A1
t
= CCH
t
+1A1
0
(17)
As spending on infrastructure and education is already high in the NMS countries, this
exercise takes domestically–…nanced expenditure 1A1
0
in the counterfactual situation
(without structural and cohesion funds) as the benchmark and only examines the impact
of the …scal transfer CCH
t
received from the EU cohesion funds (equation 17).
3.3 Implementing the interventions
The …scal transfers under the Structural and Cohesion Policy programmes amount to
additional injections of up to 3 percent of GDP for the NMS. The transfers are modelled
as lump-sum transfers between governments. Table 2 shows the main …elds of interven-
tions, their respective share in the total amount of interventions and the way each of the
interventions are captured as shocks to the model. We assume that these shares of the
…elds of interventions are constant for all the years of the payment horizon 2007-2015.
5
Herve and Holzmann (1998) criticise earlier model-based studies of structural funds for grossly ex-
aggerating the total impact because they assumed that the full Structural Fund spending is additional
to investment in the counterfactual situation TOTEXP
t
= COH
t
+c COH
t
+EXP
0
while the correct
formulation of the additionality principle is given by equation (17).
12
Table 2: Fields and Shares of Interventions
Field Share Variable to implement the shock
Infrastructure 62.0 Temporary increase in 1
G
, government investment
(via
1G
t
)
Industry&Services 9.1 Reducing …xed costs faced by …nal goods …rms (1C
Y
)
- 1.5 permanent reduction 1C
Y
(categories 13,14, 15 of the payment pro…le)
- 7.6 temporary reduction 1C
Y
RTD 10.1 Reducing the …xed costs faced by
the users of R&D products (1C¹)
- 4.5 permanent investment in R&D infrastructure
(categories 2 and 3 of the payment pro…le)
- 5.6 temporary investments
Human resources 15.3 Raising human capital and government
consumption expenditures
- 1.4 investment in high-skilled human capital (/
1
t
via
1
t
)
(category 74 of the payment pro…le)
- 4.8 educational investments in all skills (/
c
t
via
c
t
)
(categories 72 and 73 of the payment pro…le)
- 9.1 other government expenditures (G
t
)
Technical assistance 3.4 Temporary increase in government consumption
(G
t
)
Source: European Commission (DG REGIO).
The detailed categories of the payment pro…le are shown in the Annex.
Investment in public infrastructure is modelled via a temporary increase in government
investments
1G
t
. Support to industry and services-related programmes are introduced
via a temporary or (depending on the nature of the programme) permanent decrease in
…xed costs of …nal goods …rms (1C
Y
). R&D promoting spending is modelled similarly,
via decreasing the …xed costs faced by the intermediate sectors (1C
¹
) temporarily or
permanently, depending on the nature of the programme. Concerning human capital
investments we distinguish three subcategories of payments based on the detailed payment
pro…le. Around [4.8] per cent of the funds are spent on educational investments without
speci…c skill-speci…cation, and allocated in the model to all skill groups. A smaller share of
1.4 per cent is directly targeting investments in high-skilled human capital and captured
in the model as a shock to
1
t
. The remainder is accounted for as temporary increase
in government consumption (G
t
). On the basis of available data on current spending on
education (around 5 per cent on all skill-groups and 1 per cent on high-skilled in terms
13
of GDP) an estimate can be made of the additional years of schooling (increment to
c
t
)
that can be …nanced by the …scal transfers
6
. In order to account for the additional time
spent on training, we assume that the last cohort of student population stays longer in the
education system and enter into the active labour force later. Finally technical assistance
is introduced as a temporary increase in government consumption.
4 The potential macro economic impact
4.1 Simulations results
The interventions of the Structural and Cohesion Policy programmes are simulated in
the model as shocks described above and Table 3 shows the macroeconomic impact on
the new member states. There is a large increase in government spending, in particular
in government investment, and a gradual accumulation of public capital. This raises
productivity and leads to a gradual increase in output in the …nal goods sector. Growth is
further boosted by the other interventions. R&D promoting policies lower entry costs and
so reduce the pro…t requirements for intermediate producers and encourages entry of new
…rms. Higher demand for patents (ideas) increases the demand for high skilled workers and
boost innovation. Investment in training and education boost human capital and raises
output further. It takes years for these supply side e¤ects to build up and in the …rst few
years the demand e¤ects dominate. By 2020, the stock of "knowledge", or ideas, is up by
2 per cent. The incentives shift employment of high skilled workers from the …nal goods
sector to the R&D sector. In the short run, the additional spending puts upward pressure
on wages and interest rates and leads to some crowding-out, with private investment
rising proportionally less than GDP. As a result of this, in the …rst two years GDP rises
by less than the size of the demand impulse. But the output gains become gradually
larger as the supply side e¤ects become stronger, and by 2015 output rises to more than
4 per cent above baseline. These positive supply side e¤ects are permanent, as becomes
6
See the Appendix for a detailed description of the calibration of human capital accumulation.
14
apparent when looking at the e¤ects after 2015, when the programmes are assumed to be
terminated and there is no longer a demand impulse, but output is permanently higher.
Employment rises in the …rst years due to the demand expansion, but the employment
e¤ect is slightly negative afterwards. R&D enhancing policies account for signi…cant
increase in the employment share of high-skilled labour in the research sector. This leads
to some relocation of high skilled workers from production of …nal goods to the R&D
sector and an increase in the wage of high skilled workers. In the long run the total
employment e¤ect remains negative because of a negative terms of trade e¤ect.
7
Consumption of non-liquidity-constrained households is directly boosted by higher ex-
pected future income and aggregate consumption (liquidity and non-liquidity constrained)
is already in the …rst year more than 1 per cent higher. Although the business sector re-
ceives direct support from the government, this subsidy has only a small e¤ect in the
…rst years of the programmes because of crowding-out due to overproportionally higher
government spending on infrastructure and human resources which puts upward pressure
on real interest rates. In small open economies, a signi…cant share of the demand impulse
leaks abroad through higher imports. Imports are more than 2 per cent higher and the
trade balance deteriorates. The demand impulse leads to a small appreciation in the short
run but as the productivity e¤ects become stronger the e¤ective exchange rate depreci-
ates. Finally, the GDP e¤ect for the donor countries is negative, due to an increase in
their contributions to the EU budget of around 0.1 per cent of GDP, but the relatively
smaller output loss indicates a small positive spill-over e¤ect from higher growth in the
new member states.
7
To the extent that human capital investment improves the employability of the non-active population,
it could lead to an increase in the participation rate. As this e¤ect is ignored in model simulations, we
potentially underestimate the impact on employment rates.
15
Table 3: Simulated macroeconomic e¤ects
Variable 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2020
GDP 0.39 0.65 0.92 1.32 1.74 2.25 2.79 3.44 3.82 3.63 3.42
Ideas -0.00 -0.01 -0.01 0.01 0.05 0.14 0.28 0.48 0.73 1.01 1.97
TFP -0.00 -0.00 -0.00 0.00 0.00 0.01 0.02 0.04 0.06 0.09 0.17
Human cap. - low 0.01 0.04 0.06 0.09 0.12 0.17 0.21 0.27 0.32 0.35 0.32
Human cap. - medium 0.06 0.23 0.37 0.55 0.77 1.03 1.33 1.67 2.01 2.16 1.97
Human cap. - high 0.08 0.27 0.45 0.65 0.92 1.22 1.58 1.99 2.39 2.56 2.32
Employment 0.04 0.02 -0.02 -0.03 -0.06 -0.09 -0.13 -0.19 -0.32 -0.47 -0.45
- low.(…nal goods) 0.09 0.11 0.09 0.06 -0.02 -0.12 -0.25 -0.40 -0.61 -0.79 -0.61
- medium (…nal goods) 0.04 0.01 -0.03 -0.04 -0.06 -0.09 -0.12 -0.17 -0.30 -0.44 -0.44
- high (…nal goods) 0.09 0.07 0.03 -0.01 -0.11 -0.24 -0.40 -0.55 -0.77 -0.99 -0.85
- high (R&D sector) -0.25 -0.49 -0.49 -0.26 0.19 0.81 1.55 2.33 3.03 3.38 2.42
Consumption 1.19 1.94 2.20 2.39 2.60 2.84 3.08 3.30 3.44 3.44 2.93
- non.constr. hh. 1.27 2.05 2.30 2.47 2.66 2.88 3.09 3.28 3.41 3.45 3.00
- liq.constr.hh. 0.42 0.87 1.23 1.65 2.07 2.53 2.98 3.47 3.68 3.35 2.31
Investment 0.01 0.03 0.08 0.18 0.34 0.54 0.77 1.00 1.21 1.38 1.65
Exports -0.09 -0.07 -0.00 0.07 0.17 0.29 0.42 0.58 0.77 0.91 0.89
Imports 1.64 2.02 1.99 2.20 2.23 2.29 2.25 2.25 1.38 -0.35 -0.58
Real wages 0.19 0.51 0.83 1.16 1.55 1.98 2.44 2.90 3.30 3.54 3.36
Price level 0.01 -0.05 -0.16 -0.34 -0.64 -1.04 -1.55 -2.12 -2.70 -3.18 -3.96
Consumer.price level -0.07 -0.08 -0.12 -0.23 -0.42 -0.71 -1.08 -1.50 -1.91 -2.26 -3.08
Terms of trade 0.11 0.03 -0.10 -0.24 -0.41 -0.59 -0.80 -1.02 -1.28 -1.48 -1.49
REER -0.13 -0.06 0.06 0.19 0.37 0.57 0.80 1.06 1.38 1.61 1.54
Euro exchange rate -0.15 -0.17 -0.20 -0.30 -0.47 -0.73 -1.06 -1.43 -1.75 -2.02 -2.86
Nom. interest rate -0.01 0.05 0.06 0.03 -0.04 -0.13 -0.23 -0.31 -0.37 -0.42 -0.23
Real interest rate 0.02 0.14 0.20 0.29 0.33 0.35 0.32 0.31 0.20 -0.09 -0.11
In‡ation -0.00 -0.08 -0.13 -0.22 -0.34 -0.45 -0.54 -0.60 -0.58 -0.41 -0.13
Cons. in‡ation -0.07 -0.02 -0.06 -0.13 -0.23 -0.32 -0.40 -0.43 -0.40 -0.32 -0.14
R&D intensity -0.01 -0.01 -0.01 -0.01 0.00 0.01 0.03 0.04 0.06 0.07 0.05
Lab. productivity 0.52 0.97 1.42 2.05 2.73 3.55 4.44 5.52 6.31 6.24 5.90
Employment rate 0.03 0.01 -0.01 -0.02 -0.04 -0.06 -0.09 -0.12 -0.21 -0.31 -0.30
- low 0.04 0.05 0.04 0.02 -0.02 -0.08 -0.15 -0.23 -0.34 -0.44 -0.43
- medium 0.03 0.00 -0.02 -0.03 -0.04 -0.05 -0.07 -0.10 -0.19 -0.28 -0.26
- high 0.02 -0.02 -0.05 -0.04 -0.04 -0.03 -0.02 0.01 -0.03 -0.12 -0.12
Coh. payments. (%GDP) 1.00 0.89 0.89 1.29 1.48 1.78 1.98 2.37 1.78 0.00 0.00
Gov.debt (%GDP) -0.32 -0.65 -0.88 -1.14 -1.36 -1.59 -1.80 -2.02 -2.01 -1.52 -0.26
Gov.balance (%GDP) 0.28 0.21 0.16 0.19 0.18 0.20 0.20 0.20 -0.01 -0.35 -0.16
Trade balance (%GDP) -0.96 -1.23 -1.25 -1.41 -1.47 -1.56 -1.58 -1.62 -1.16 -0.20 -0.08
Euro-area GDP 0.01 -0.01 -0.03 -0.03 -0.04 -0.04 -0.05 -0.04 -0.05 -0.09 -0.07
Note: Percentage (top half) and absolute (bottom half) di¤erences from baseline
16
4.2 Sensitivity analysis
The results are sensitive to the assumption on the marginal product of public capital.
There is much uncertainty about the productive impact of infrastructure investment and
econometric studies show large variation in estimates, depending how care is taken of
common trends, missing variables, simultaneity bias and reverse causation. Many of the
early empirical estimates have been dismissed as implausibly high
8
. The benchmark
assumption in the model is that the rate of return on public capital equals that of private
capital (Gramlich (1994), p.1187). However, for some of the infrastructure interventions
(e.g. environmental protection and development of cultural and social infrastructure) it
could be argued that this assumption is too optimistic. Figure 2 plots the GDP e¤ect
of the programmes under three alternative scenarios relative to our baseline assumption
for c
G
(continuous middle line in the …gure). The sensitivity analysis shows substantial
variations both in the peak as well as in the long-run GDP e¤ects, of around a quarter
of the benchmark e¤ect. The projected output gain in 2015 can be as high as 5 per cent
under a more favourable assumption of c
G
= 0.15, while it could be as low as 3 per cent
when c
G
= 0.05. Depending on the magnitude of the marginal productivity of capital, it
can take from 1 to 6 years till the supply side improvements raise the level of GDP above
the level of the direct demand injection from the programmes.
The results are also highly sensitive to how we interpret the interventions on human
capital formation. Total interventions classi…ed as human capital investment in the Co-
hesion and Structural Funds account for 15 per cent of total spending but this category
covers a wide range of measures, including e.g. various measures improving the social
inclusion of less-favoured groups.(see annex). It is not a priori clear whether all these
measures improve skill e¢ciencies as much as assumed in the model simulations. A sen-
sitivity analysis to the speci…c assumptions concerning this category shows the impact
these assumptions have on the overall results. Our benchmark scenario assumes that only
those categories that are explicitly labelled as improving human capital in the programmes
8
For an overview of the literature and the econometric problems, see e.g. the surveys by Gramlich
(1994).
17
Figure 2: GDP e¤ects under di¤erent assumptions on the marginal productivity of public capital
(% di¤erence from baseline)
0
1
2
3
4
5
6
7
2
0
0
7
Q
1
2
0
0
7
Q
3
2
0
0
8
Q
1
2
0
0
8
Q
3
2
0
0
9
Q
1
2
0
0
9
Q
3
2
0
1
0
Q
1
2
0
1
0
Q
3
2
0
1
1
Q
1
2
0
1
1
Q
3
2
0
1
2
Q
1
2
0
1
2
Q
3
2
0
1
3
Q
1
2
0
1
3
Q
3
2
0
1
4
Q
1
2
0
1
4
Q
3
2
0
1
5
Q
1
2
0
1
5
Q
3
2
0
1
6
Q
1
2
0
1
6
Q
3
2
0
1
7
Q
1
2
0
1
7
Q
3
2
0
1
8
Q
1
2
0
1
8
Q
3
2
0
1
9
Q
1
2
0
1
9
Q
3
2
0
2
0
Q
1
2
0
2
0
Q
3
quarters
%
0
1
2
3
4
5
6
7
Net-transfers received (% GDP) alphag=0.10, benchmark alphag=0.05 alphag=0.15
(i.e. categories 72,73 and 74, 6.2 per cent of total spending) are a¤ecting human capital
accumulation in the model. Our alternative scenario shown in Figure 2 plots the GDP
e¤ect if all spending on human resources (15.3 per cent of toatl spending) is assumed
to be educational investments to improve human capital at each skill-level. While this
assumption may be too optimistic given the social nature of some of this spending, it
gives a clear upper bound of the potential impact this spending can have on output.
18
Figure 3: GDP e¤ects under more optimistic scenario on human capital investments (% di¤er-
ence from baseline)
0
1
2
3
4
5
6
7
2
0
0
7
Q
1
2
0
0
7
Q
3
2
0
0
8
Q
1
2
0
0
8
Q
3
2
0
0
9
Q
1
2
0
0
9
Q
3
2
0
1
0
Q
1
2
0
1
0
Q
3
2
0
1
1
Q
1
2
0
1
1
Q
3
2
0
1
2
Q
1
2
0
1
2
Q
3
2
0
1
3
Q
1
2
0
1
3
Q
3
2
0
1
4
Q
1
2
0
1
4
Q
3
2
0
1
5
Q
1
2
0
1
5
Q
3
2
0
1
6
Q
1
2
0
1
6
Q
3
2
0
1
7
Q
1
2
0
1
7
Q
3
2
0
1
8
Q
1
2
0
1
8
Q
3
2
0
1
9
Q
1
2
0
1
9
Q
3
2
0
2
0
Q
1
2
0
2
0
Q
3
quarters
%
0
1
2
3
4
5
6
7
Net-transfers received (% GDP) Human capital, benchmark Human capital, upper-bound
5 Conclusions
This paper descibes how a DSGE model with endogenous growth can be used for an
ex-ante (prospective) evaluation of the potential impact of EU Cohesions and Structural
Funds. The model simulates the impact of increased public infrastructure investment and
captures the productivity enhancing e¤ects of this. The model incorporates endogenous
human capital accumulation and simulates the e¤ects of policies promoting vocational
education and training on skill e¢ciencies. And the model applies the Jones (1995) semi-
endogenous growth framework to explicitly model the underlying development of R&D,
which allows us to analyse the impact of R&D promoting policies on growth.
An important caveat for model based analyses like these is that they only tell us
something about the potential impact of EU Funds, assuming an e¢cient and optimal
use. There are strong reasons to expect at least part of the total spending to be diverted
to sub-optimal use (Herve and Holzmann (1998)), and in that sense our results give an
19
upper bound of the likely e¤ects. It should also be recognised that the objective of EU
Cohesion Policy is economic and social cohesion, and hence wider than economic growth
per se. Promotion of tourism, culture, urban and rural regeneration are equally important
policy objectives.
A senstitivity analysis shows how our results depend on assumptions related to the
productivity of various forms of infrastructure investment and that of policies promoting
vocational education and training. Much uncertainty exists on these productivity e¤ects
and more micro analysis is clearly needed to shed light on these e¤ects. Future research
should also focus on how policies promoting social inclusion increase participation in the
labour force and how these e¤ects can be incorporated in the model.
References
Barro, R. J. and Sala-I-Martin, X. (1995). Economic Growth. McGraw Hill, New York.
Bottazzi, L. and Peri, G. (2007). The international dynamics of R&D and innovation in
the long run and in the short run. The Economics Journal, 117(3):486–511.
Bradley, J., Untiedt, G., and Mitze, T. (2007). Analysis of the impact of cohe-
sion policy. a note explaining the HERMIN-based simulations. Technical note,
http://ec.europa.eu/regionalpolicy/sources/docgener/evaluation/pdf/hermin07.pdf.
Caselli, F. and Coleman, W. J. (2006). The world technology frontier. American Economic
Review, 96(3):499–522.
Dixit, A. K. and Stiglitz, J. E. (1977). Monopolistic competition and optimum product
diversity. American Economic Review, 67(3):297–308.
Djankov, S., Porta, R. L., Lopez-De-Silanes, F., and Shleifer, A. (2002). The regulation
of entry. The Quarterly Journal of Economics, 117(1):1–37.
Ederveen, S., Groot, H., and Nahuis, R. (2006). Fertile soil for structural funds? a
panel data analysis of the conditional e¤ectiveness of european cohesion policy. Kyklos,
Blackwell Publishing, 59(1):pp. 17–42.
Ederveen, S., Groter, J., de Mooji, R., and Nahuis, R. (2002). Funds and games: The eco-
nomics of european cohesion policy. Special Publication 41, CPB Netherlands’ Bureau
for Economic Policy Analysis, The Hague.
Gramlich, E. (1994). Infrastructure investment: A review essay. Journal of Economic
Literature, 32(3):1176–1196.
20
Gros, D. (2008). How to achieve a better budget for the european union? CEPS Working
Document 289, Center for European Policy Studies, Brussels.
Herve, Y. and Holzmann, R. (1998). Fiscal Transfers and Economic Convergence in the
EU : An Analysis of Absorption Problems and an Evaluation of the Literature. Nomos
Verlagsgesellschaft, Baden-Baden.
Jones, C. I. (1995). R&D-based models of economic growth. Journal of Political Economy,
103(4):759–84.
Jones, C. I. (2002). Source of U.S. economic growth in a world of ideas. American
Economic Review, 92(1):220–239.
Katz, L. F. and Murphy, K. M. (1992). Changes in relative wages, 1963-1987: Supply
and demand factors. Quarterly Journal of Economics, 107(1):35–78.
Mincer, J. (1974). Schooling, Experience, and Earnings. Columbia University Press, New
York.
OECD (2006). Education at Glance. OECD, Paris.
Pessoa, A. (2005). Ideas driven growth: The OECD evidence. Portuguese Economic
Journal, 4(1):46–67.
Ratto, M., Roeger, W., and in ’t Veld, J. (2008). QUEST III: An estimated DSGE model
of the euro area with …scal and monetary policy. Economic Modelling (forthcoming),
(also published as: European Economy Economic Paper no.335 , European Commission
Directorate-General for Economic and Financial A¤airs, Brussels, July 2008).
Roeger, W., Varga, J., and in ’t Veld, J. (2008). Structural reforms in the EU: A
simulation-based analysis using the QUEST model with endogenous growth. Euro-
pean Economy Economic Paper 351, European Commission Directorate-General for
Economic and Financial A¤airs, Brussels.
Warda, J. (2006). Tax treatment of business investments in intellectual assets: An inter-
national comparison. OECD Science, Technology and Industry Working Papers 2006/4,
OECD, Paris.
21
Appendix: Calibrating the parameters of knowledge
production, intermediate goods production and hu-
man capital accumulation
Our semi-endogenous model builds on Jones (1995) version of R&D modelling, but we
account for international R&D spillovers following Bottazzi and Peri (2007). The model
economy is populated by households, …nal and intermediate goods producing …rms, a
research industry, a monetary and a …scal authority. Final good producers use a compos-
ite of intermediate goods and three types of labour - (low-, medium-, and high-skilled).
Households buy the patents of designs produced by the R&D sector and license them
to the intermediate goods producing …rms. The intermediate sector consists of monop-
olistically competitive …rms which have entered the market by licensing a design from
domestic households (at rate of i
¹,t
) and by making an initial payment 1C
¹
to overcome
administrative entry barriers. Capital inputs are also rented from the household sector
for a rental rate of i
1
t
. Firms which have acquired a design can transform each unit of
capital into a single unit of an intermediate input. The production of new designs takes
place in research labs, employing high skilled labour and making use of the existing stock
of ideas. The driving equation system of the semi-endogenous technological change can
be summarized as
¹
t
= i¹
v
t1
¹
ç
t1
1
A
¹,t
(a)
1 +q
¹
= (1 +q
a
)
A
1ç¸
(b)
` 1
¹,t
¹
t
= n
1
1
¹,t
(c)
:di =
1
¹,t
¹
t
1
Y
1
t
(d)
i
¹,t
1
¹,t
+:
t
1C
¹
= :
t
. n/c:c :
t
=
_
1
o
1
_
r
t
(e)
i
¹
=
(1 t
¹
)(i
t
:
¹
t+1
+o
¹
) t
1
o
¹
(1 t
1
)
+:j
¹
t
(f)
1
t
= ¹
t
r
t
(g)
The …rst equation is the spillover-augmented version of Jones (1995) R&D production.
This form of R&D equation accounts for international spillovers almost identically to the
speci…cation of Bottazzi and Peri (2007). Innovation corresponds to the discovery of
a new variety of producer durables that provides an alternative way of producing the
…nal good. The R&D sector hires high-skilled labour (1
¹
) and makes use of the existing
stock of domestic and foreign ideas to generate new designs (¹
t
). Parameters ¬ and c
22
measure the foreign and domestic spillover e¤ects from the aggregate international and
domestic stock of knowledge (¹

and ¹) respectively. Negative value for these parameters
can be interpreted as the "…shing out" e¤ect, i.e. when innovation decreases with the
level of knowledge, while positive values refer to the "standing on shoulders" e¤ect and
imply positive research spillovers. Note that c = 1 would give back the strong scale
e¤ect feature of fully endogenous growth models with respect to the domestic level of
knowledge. Parameter i can be interpreted as total factor e¢ciency of R&D production,
while ` measures the elasticity of R&D production on the number of researchers. The
international stock of knowledge grows exogenously at rate q
¹
u. We assume that the
R&D sector is operated by a research institute which employs high skilled labour at
their market wage n
1
. Equation (b) states the balanced-growth relationship between
the growth of ideas q
¹
(= q
¹
u) and population q
a
, equation (c) shows the …rst order
condition of R&D production, equation (d) is the de…nition of R&D-intensity: total R&D
expenditure of the intermediate sector in percentage of GDP. Equation (e) states the free-
entry condition between the pro…t of the intermediate sector (:
t
), and the per unit price
of R&D inventions (1
¹
) and the …xed (entry) cost 1C
¹
. Decreasing entry costs lowers the
pro…ts requirement for intermediate producers and thus increases entry of new …rms and
the demand for patents. Equation (f) de…nes the rental rate of intangible capital which
takes into account that households pay income tax at rate t
1
t
on the period return of
intangibles and they receive tax subsidies at rate t
¹
. Since one unit of capital is used to
produce one unit of intermediate good (r
t
), equation (g) states the identity between the
total intermediate goods production and physical capital under symmetric equilibrium.
A. 1. R&D production
Although we do not have direct estimates of i, ¬, c and `, we can use the existing
literature and the model restrictions to get calibrated values for them. Data on the R&D
share of labour ( 1
¹,t
) and on the R&D intensity
_
1
/,I
¹
T
I
1
Y
Y
I
_
is obtained from EUROSTAT,
the values of q
¹
and q
a
are given in our baseline model
9
. These values together with the
restrictions of the balanced growth dynamics and the other variables of the baseline pin
down ` and 1
¹
. In order to set c and ¬ in the …rst step we express the sum of these
two parameters from equation (b). In the second step we use the estimated long-term
relationship between ` and ¸ from Bottazzi and Peri (2007) to approximate ¬ separately.
The authors do not estimate directly c and ¬, however their estimated cointegration
vector contains two coe¢cients j and ¸, satisfying the following theoretical restrictions
between the long-term coe¢cients of `, c and ¬:
j =
`
|cajtcvn
1 c
|cajtcvn
and
¸ =
¬
|cajtcvn
1 c
|cajtcvn
.
The estimated values for these two coe¢cients show fairly big variations under the di¤er-
ent regressions, and it might be inadequate to apply these long-term coe¢cients on our
9
Pessoa (2005) provides estimates for the growth of patents or ideas in various OECD countries at
an average of g
A
= 0:057. The population growth g
n
is obtained from EUKLEMS potential output
calculations.
23
"contemporary" speci…cation. However the ratios of these two coe¢cients
_
¸
j
=
¬
|cajtcvn
`
|cajtcvn
_
vary less, furthermore, imposing the ratio of the long-term parameters instead of their
exact values is also less restrictive. To approximate our ¬ for the EU27, we use the ratio
of these parameters from the speci…cation in which the authors omitted the US from
their regressions
10
. In the last step we subtract this value from the sum of c and ¬ as
we calculated from equation (b) earlier. Finally, we normalize the stock of domestic and
foreign ideas to one and therefore the values for i and o can be obtained from expressions
(a) and (e).
A. 2. Intermediate goods production
The calibration of the parameters in intermediate goods production relies on the entry
costs estimations of Djankov et al. (2002), and the estimations for R&D related subsidies
(t
¹
) of Warda (2006). Given that we normalized the stock of domestic ideas to one (¹
t
),
equation (g) pins down the per …rm quantity of intermediate goods production. The
pro…t of a representative intermediate …rm is determined by its production and the net
mark-up of the sector
11
. All other variables given, the arbitrage equation (e) determines
the rental rate of intangible capital, i
¹
t
. The B-indices published in Warda (2006) can be
applied to calibrate t
¹
and t
1
. Finally, we use the de…nition of equation (f) to obtain as
residual the calibrated approximation of the risk-premium on intangibles, :j
¹
t
.
A. 3. Human capital accumulation
Labour force is disaggregated into three skill-groups: low-, medium- and high-skilled
labour. The CES-aggregate for labour have the following form:
1
Y,t
=
_
:
1
¬
1
1
_
/
1
t
1
1
t
_
¬
1
1
¬
1
+:
1
¬
1
A
_
/
A
t
1
A
t
_
¬
1
1
¬
1
+:
1
¬
1
1,Y
_
/
1
t
1
1Y
t
_
¬
1
1
¬
1
_
¬
1
¬
1
1
.
where the subscripts denote the skill-groups (low-1, medium-` and high-H), :
c
is
the population share of labour-force in subgroup :, 1
c
denotes the employment rate of
population :, /
c
t
is the skill-speci…c e¢ciency unit of labour, and o
1
is the elasticity of
substitution between di¤erent labour types. Note that high-skilled labour in the …nal
goods sector is the total high-skill employment minus the high-skilled labour working
for the R&D sector (1
¹,t
). The calibration is mostly based on EUROSTAT and OECD
data. Data on skill-speci…c population shares, participation rates and wage-premiums
are obtained from the Labour Force Survey and Science and Technology databases of
EUROSTAT. The elasticity of substitution between di¤erent labour types (o
1
) is one
of the major issue addressed in the labour-economics literature. We follow Caselli and
Coleman (2006) which analyzed the cross-country di¤erences of the aggregate production
function when skilled and unskilled labour are imperfect substitutes. The authors argue in
10
The full sample consists of …fteen OECD countries including the US and ten member states of the
European Union.
11
We use the net mark-up of the manufacturing sector calculated in EUKLEMS to obtain , the inverse
of the gross mark-up in the intermediate sector.
24
favour of using the Katz and Murphy (1992) estimate of 1.4. We normalize the e¢ciency
of low-skilled at 1 the other e¢ciency units are restricted by the labour demand equations
which imply the following relationship between wages, labour-types and e¢ciency units:
/
A
t
=
_
n
A
n
1
_
¬
1
¬
1
1
_
:
A
1
A
:
1
1
1
_ 1
¬
11
/
1
t
. c:d /
1
t
=
_
n
1
n
A
_
¬
1
¬
1
1
_
:
1
1
1
:
A
1
A
_ 1
¬
11
/
A
t
.
In the next step we adapt Jones (2002) into a disaggregated skill-structure and impose
that the functional form of /
c
t
= /
c
c
ç
s
I
describes the evolution of skill-speci…c human
capital. In line with Jones (2002), we …x the return to schooling parameter of · at 0.07.
The number of school years,
c
t
for the respective skill-groups are obtained from OECD
(2006). For simulation purposes, the participation in trainings can be interpreted as an
addition to the years of schooling with a depreciation according to the exit rate of working
age population, i.e.:

c
t
=
c
+|
c,T1
t
. n/c:c |
c,T1
t
= (1 ¸
c
) |
c,T1
t1
+
c,T1
t
.
where for each skill-group :,
c
is the average number of years of schooling in the
regular education system, |
c,T1
t
is the year equivalent of the average time spent in training
in period t, ¸
c
is the exit-rate of the working age population, and
c,T1
t
is the average year-
equivalent of training in period t. Finally, in the baseline we set the variables of training
|
c,T1
t
and
c,T1
t
to zero and given the years of schooling from OECD (2006) we can compute
/
c
from the de…nition of e¢ciency. In order to simulate the educational investments in
human capital we increase the years of schooling (
c
t
) for the respective skill-groups by
the additional years of schooling that can be …nanced from the …scal transfers (shock to

c,T1
t
).
A. 4. Population dynamics
Denote by `o
c
(t) and `\
c
(t), : 2 f1. `. Hg respectively the number of students and
workers in skill-group :, in year t.
The evolution of school population can be written as
`o
c
(t) = (1 ¸
c
+/
c
) `o
c
(t 1).
and for the working population
`\
c
(t) = (1 ¸
c
) `\
c
(t 1) +¸
c
`o
c
(t 1).
where ¸
c
is the exit rate from student skill cohort :, /
c
is the birth-rate of skill group
:, and ¸
c
is the exit rate from working age population. The exit rate is the inverse of
average duration spent in a skill-group. For example, if the duration of education for high
skilled is 20 years (80 quarters), then ¸
c
= 1,80.
The parameters /
c
, ¸
c
, and ¸
c
determine the evolution of the (inverse) dependency
ratios
11111¹11
c
(t) =
`\
c
(t)
`o
c
(t)
and the growth rate of working age population (qjojn
c
(t)). Combining the dynamic
25
equations of the student and working age population we get
11111¹11
c
(t) =
1 ¸
c
1 ¸
c
+/
c
11111¹11
c
(t 1) +
¸
c
1 ¸
c
+/
c
.
therefore the growth rate of the working age population is given by
`\
c
(t)
`\
c
(t 1)
=
¸
c
11111¹11
c
(t 1)
¸
c
.
Finally, by adding the de…nitions
:/
c
(t) =
`\
c
(t)

c2f1,A,1g
`\
c
(t)
we obtain the evolution of the skill population shares. For the calibration, we use
the education statistics of OECD (2006) and data on the skill-distribution of working age
population from EUROSTAT.
26
Annex: Detailed payment pro…le categories
Table A1. Fields of interventions, detailed tables
Code Category Priority themes % of total
Research and technological development (R&TD),
innovation and entrepreneurship
01 RTD R&TD activities in research centres 1.21
02 RTD R&TD infrastructure and centres of competence 2.78
in a speci…c technology
03 RTD Technology transfer and improvement of cooperation 1.32
networks between small businesses (SMEs),
between these and other businesses and universities,
postsecondary education establishments of all kinds
regional authorities, research centres and scienti…c
and technological poles
(scienti…c and technological parks, technopoles, etc).
04 RTD Assistance to R&TD, particularly in SMEs 0.73
(including access to R&TD services in research centres)
05 S Advanced support services for …rms and groups of …rms 1.17
06 I Assistance to SMEs for the promotion of 0.46
environmentally-friendly products and
production processes (introduction of e¤ective environment
managing system, adoption and use of pollution
prevention technologies, integration of clean
technologies into …rm production
07 RTD Investment in …rms directly linked to research and innovation 2.85
(innovative technologies, establishment of new …rms
by universities, existing R&TD centres and …rms, etc.)
08 I Other investment in …rms 2.88
09 RTD Other measures to stimulate research and innovation 1.13
and entrepreneurship in SMEs
Information society
10 INF Telephone infrastructures (including broadband networks) 0.75
11 INF Information and communication technologies 1.05
(access, security, interoperability, risk-prevention,
research, innovation, e-content, etc.)
12 INF Information and communication technologies (TEN-ICT) 0.08
13 S Services and applications for citizens 1.55
(e-health, e-government, e-learning, e-inclusion, etc.)
14 S Services and applications for SMEs 0.65
(e-commerce, education and training, networking, etc.)
15 S Other measures for improving access to 0.45
and e¢cient use of ICT by SMEs
27
Code Category Priority themes % of total
Transport
16 INF Railways 0.94
17 INF Railways (TEN-T) 6.96
18 INF Mobile rail assets 0.29
19 INF Mobile rail assets (TEN-T) 0.40
20 INF Motorways 1.87
21 INF Motorways (TEN-T) 8.37
22 INF National roads 3.47
23 INF Regional/local roads 3.69
24 INF Cycle tracks 0.22
25 INF Urban transport 0.78
26 INF Multimodal transport 0.23
27 INF Multimodal transport (TEN-T) 0.19
28 INF Intelligent transport systems 0.43
29 INF Airports 0.54
30 INF Ports 0.58
31 INF Inland waterways (regional and local) 0.06
32 INF Inland waterways (TEN-T) 0.27
Energy
33 I Electricity 0.08
34 INF Electricity (TEN-E) 0.15
35 I Natural gas 0.30
36 INF Natural gas (TEN-E) 0.14
37 I Petroleum products 0.09
38 INF Petroleum products (TEN-E) 0.00
39 I Renewable energy: wind 0.25
40 I Renewable energy: solar 0.18
41 I Renewable energy: biomass 0.52
42 I Renewable energy: hydroelectric, geothermal and other 0.25
43 I Energy e¢ciency, co-generation, energy management 1.28
28
Code Category Priority themes % of total
Environmental protection and risk prevention
44 INF Management of household and industrial waste 2.60
45 INF Management and distribution of water (drink water) 2.52
46 INF Water treatment (waste water) 5.37
47 INF Air quality 0.48
48 INF Integrated prevention and pollution control 0.23
49 INF Mitigation and adaption to climate change 0.01
50 INF Rehabilitation of industrial sites and contaminated land 1.17
51 INF Promotion of biodiversity and nature protection 0.77
(including Natura 2000)
52 INF Promotion of clean urban transport 2.56
53 INF Risk prevention (including the drafting and 1.57
implementation of plans and measures to
prevent and manage natural and technological risks)
54 INF Other measures to preserve the environment and prevent risks 0.45
Tourism
55 S Promotion of natural assets 0.31
56 INF Protection and development of natural heritage 0.35
57 S Other assistance to improve tourist services 1.13
Culture
58 INF Protection and preservation of the cultural heritage 0.82
59 INF Development of cultural infrastructure 0.78
60 S Other assistance to improve cultural services 0.11
Urban and rural regeneration
61 INF Integrated projects for urban and rural regeneration 2.52
29
Code Category Priority themes % of total
Increasing the adaptability of workers and …rms,
enterprises and entrepreneurs
62 HC Development of life-long learning systems and strategies in …rms; 1.56
training and services for employees to step up their adaptability
to change; promoting entrepreneurship and innovation
63 HC Design and dissemination of innovative and 0.50
more productive ways of organising work
64 HC Development of speci…c services for employment, 0.75
training and support in connection with restructuring of sectors
and …rms, and development of systems for anticipating economic
changes and future requirements in terms of jobs and skills
Improving access to employment and sustainability
65 HC Modernisation and strengthening labour market institutions 0.57
66 HC Implementing active and preventive measures on the labour market 1.76
67 HC Measures encouraging active ageing and prolonging working lives 0.25
68 HC Support for self-employment and business start-up 0.43
69 HC Measures to improve access to employment and increase 0.37
sustainable participation and progress of women in employment
to reduce gender-based segregation in the labour market,
and to reconcile work and private life, such as facilitating
access to childcare and care for dependent persons
70 HC Speci…c action to increase migrants’ participation in employment 0.03
and thereby strengthen their social integration
Improving the social inclusion of less-favoured persons
71 HC Pathways to integration and re-entry into employment 1.50
for disadvantaged people; combating discrimination
in accessing and progressing in the labour market
and promoting acceptance of diversity at the workplace
Improving human capital
72 HC Design, introduction and implementation of reforms in education and 2.86
training systems in order to develop employability, improving the
labour market relevance of initial and vocational education and
training, updating skills of training personnel with a view to
innovation and a knowledge based economy
73 HC Measures to increase participation in education and training 1.95
throughout the life-cycle, including through action to achieve
a reduction in early school leaving, gender-based segregation of
subjects and increased access to and
quality of initial vocational and tertiary education and training
74 HC Developing human potential in the …eld of research and innovation, 1.36
in particular through post-graduate studies and
training of researchers, and networking activities
between universities, research centres and businesses
30
Code Category Priority themes % of total
Investment in social infrastructure
75 INF Education infrastructure 2.32
76 INF Health infrastructure 2.19
77 INF Childcare infrastructure 0.19
78 INF Housing infrastructure 0.39
79 INF Other social infrastructure 0.85
Mobilisation for reforms
in the …elds of employment and inclusion
80 HC Promoting the partnerships, pacts and 0.16
initiatives through the networking of relevant stakeholders
Strengthening institutional capacity at
national, regional and local level
81 HC Mechanism for improving good policy 1.24
and programme design, monitoring and evaluation
Reduction of additional costs hindering
the outermost regions development
82 S Compensation of any additional costs due to 0.00
accessibility de…cit and territorial fragmentation
83 S Speci…c action addressed to compensate 0.00
additional costs due to size market factors
84 I Support to compensate additional costs due to 0.00
climate conditions and relief di¢culties
Technical assistance
85 Preparation, implementation, monitoring and inspection 2.71
86 Evaluation and studies; information and communication 0.74
31

Economic Papers are written by the Staff of the Directorate-General for Economic and Financial Affairs, or by experts working in association with them. The Papers are intended to increase awareness of the technical work being done by staff and to seek comments and suggestions for further analysis. The views expressed are the author’s alone and do not necessarily correspond to those of the European Commission. Comments and enquiries should be addressed to: European Commission Directorate-General for Economic and Financial Affairs Publications B-1049 Brussels Belgium E-mail: Ecfin-Info@ec.europa.eu

This paper exists in English only and can be downloaded from the website http://ec.europa.eu/economy_finance/publications A great deal of additional information is available on the Internet. It can be accessed through the Europa server (http://europa.eu )

KC-AI-09-371-EN-N ISBN 978-92-79-11182-2 ISSN 1725-3187 DOI 10.2765/2997

© European Communities, 2009

A Model-based Assessment of the Macroeconomic Impact of EU Structural Funds on the New Member States
Janos Varga and Jan in ’ Veld t March 9, 2009

Abstract

This paper gives a model-based analysis of the potential macro-economic impact of European Union Structural and Cohesion Funds payments on the economies of the new member states. The model used is a four-region DSGE model with human capital accumulation and endogenous technological change. The framework that we adopt is the Jones (2005) extension of the endogenous growth model, which uses a variety approach for modelling knowledge investment. The EU funds average around 1.5 percent of GDP and are used for investment in infrastructure, human capital and R&D. The model simulations show this can lead to signi…cant gains in output, both in the short as well as in the long run. JEL Classi…cation: E62, H50, O11, O41 Keywords: Fiscal transfers, Structural Funds, public investment, DSGE modelling.

We like to thank Dalia Grigonyte for providing us with the dataset. The views expressed in this paper are those of the authors and should not be attributed to the European Commission. Correspondence: Janos Varga, European Commission, DG ECFIN Economic and Financial A¤airs, B-1049 Brussels, Belgium, email: janos.varga@ec.europa.eu ; Jan in ’ Veld, European Commission, DG ECFIN Economic t and Financial A¤airs, B-1049 Brussels, Belgium, email: jan.intveld@ec.europa.eu

Bradley et al. empirical evidence on the positive impact of Structural 1 . Model based studies. these studies have often been based on macroeconometric models in which the demand side is essentially keynesian in nature and no-crowding-out appears (e. The funds are also seen as an expression of solidarity between member states and a commitment to redistribution of incomes considered necessary for an e¤ective functioning of the Union’ political institutions (Gros (2008)). accounting for 34 per cent of total appropriations for commitments. ex-post evaluation studies of large EU transfers in the past generally give only mixed support for EU Cohesion Policy. However. after agricultural spending. Many of the assisted regions have remained relatively poor and growth regressions augmented with Structural Fund variables show no signi…cant impact from these transfers (see Ederveen et al. (2007)). the second largest item in the current EU budget. frequently used for the purposes of ex-ante evaluation. and with EU enlargement in 2004 there has been a sharp increase in EU Structural Funds.g. usually show larger impacts of the demand stimulus as well as positive supply-side e¤ects on GDP in the long run.1 Introduction The European Union’ Cohesion Policy provides a framework for large …scal transfers from s the richer EU Member States to the countries and regions that lag behind in terms of income per capita. although special ad hoc arrangements exist. s Economic theory predicts unambiguous bene…ts from investment in infrastructure and human capital and there is empirical evidence supporting this. The Cohesion and Structural Funds are. On the whole. it was argued that poorer member states had to be compensated for exposing their weaker economies to more competition in the internal market. enhancing a faster catch-up process of the less developed member states. However. For eligibility for Objective One status the region’ GDP per capita must be less than 75 per s cent of the EU average. and designed to increase economic and social cohesion among member states. At the time of the creation of the Single Market programme. (2002) and Ederveen et al. The resources are targeted on investment in physical and human capital. (2006)).

(2008)) and is particularly suitable for an evaluation of the type of structural policies that form the core of Structural Funds interventions.1 The model we use in this study is based on a four region dynamic general equilibrium model with human capital accumulation and endogenous technological change. it is important to examine whether Structural Funds can promote growth and enhance a faster catching up of the less developed member states. improving social inclusion of less-favoured persons. This paper uses the QUEST III model to evaluate the potential impact of Structural and Cohesion Fund programmes for the new member states for the period 2007-2013. While these policies undoubtedly promote social cohesion. 1 2 . Hence. A cursory look at the …elds of interventions of CSF shows a wide range of policies promoting environmental protection and risk prevention. rent-seeking may be very high.Funds on convergence of poorer regions remains inconclusive. It has been used extensively for the analysis of structural reforms in the EU (the Lisbon Strategy for Growth and Jobs) (see Roeger et al. While it should be clear that an ex-ante model-based evaluation cannot provide a conclusive answer to the question whether structural funds will accelerate the economic convergence of poorer regions. they are not directly enhancing the growth potential of an economy. at least not in a unequivocal and directly measurable way. It should be recognised that the objective of cohesion policy is much wider than economic growth per se. culture. The model incorporates productive public investment that captures the productivity-enhancing e¤ects of infrastructure investment. Herve and Holzmann (1998) list many factors that could lead to sub-optimal use of …scal transfers and argue that the scope for problems related to e. tourism. Nevertheless.g. The model belongs to the new class of micro-founded DSGE models that are now widely used in economic policy institutions. an analysis based on a model that incorporates the channels through which these policies can promote growth can provide an estimate of the potential impact. urban and rural regeneration. economic convergence is the ultimate objective of cohesion policy and the eligibility criterium for Objective One status is clearly de…ned in terms of income per capita. The model employs the product variety framework proposed by Dixit and Stiglitz (1977) and applies the Jones In a comprehensive study of absorption problems. assuming an optimal use of the funds.

But because past experience in previous programme periods have shown considerable delays in payments. typically continuing for up to two more years. more than 60% of the total budget for most NMS. The model results depend crucially on 1. while investments in R&D and human capital are the second or third largest categories. The simulation results and a sensitivity analysis with respect to these assumptions are discussed in Section 4. on how the skill e¢ ciencies are a¤ected by human capital investment. we calculate in Table 1 the proposed s annual payment pro…le in terms of GDP for 2007-2015.9 billion euros (in 2008 prices). The paper is organized as follows. Finally. assuming an in‡ ation correction of 2 per cent per year. The next section brie‡ discusses the Structural and y Cohesion Funds programmes and the …scal transfers involved. The …elds of interventions are divided into …ve main categories (plus an additional technical assistance category). the assumptions on the productive impact of additional public capital and 2. Section 5 concludes. As the …gure shows. the proposed payments are spread over a 9 year period lasting till 2015. The …elds of intervention cover a wide range of 3 . 2 The Structural and Cohesion Funds programme 2007-2013 For the period 2007 to 2013. infrastructure investment receives the largest share of funds. The endogenous modelling of R&D allows us to analyse the impact of R&D promoting policies on growth and the endogeneity of human capital accumulation can capture the e¤ects of policies promoting vocational education and training. On the basis of payment pro…les in programming prices.(1995) semi-endogenous growth framework to explicitly model the underlying development of R&D. and using the Commission’ nominal GDP projections for 2008-15. Section 3 describes those features of the model which enable us to carry out the impact assessment of the …scal transfers. Structural and Cohesion Funds programmes for the New Member States amount to a total budget of 173. Figure 1 shows the distribution of the total budget for each of the New Member States.

3 0.9 3.6 1. The payment pro…le in programming prices was provided by DG REGIO with an assumption of 2% in‡ ation.7 2.8 0.3 1.1 0.7 1. (2008).9 1.0 0.1 1.9 1.4 1.8 1.3 Slovenia 0.0 1.9 0.9 0.7 1. the Euro area and non-Euro area old member states and the rest of the world.2 1.9 2.8 1. details of which are shown in the annex (Table A1).2 0.9 1.2 2.3 0.1 1.1 1.7 2.9 2.0 Lithuania 1.9 1. This section describes the main channels through which the model captures the impact of the di¤erent interventions on growth.0 2.4 0.2 2. The model consists of four regions: the block of new member states.1 1.3 1.5 1.4 2.4 0.8 Slovakia 1.5 0.7 2.8 1.2 1.4 1.0 2.4 0.1 1.2 1. GDP projections are obtained from the AMECO database and DG ECFIN’ potential growth calculations.2 1.3 1.2 2.2 0.1 1.3 0.7 1.0 1.4 1.7 0.9 Hungary 1.4 1.0 1. 3.6 0.4 Poland 1.5 1.8 2.2 0.7 0.8 1.0 1.1 0.9 0.9 1. s policy programmes.8 2.9 2. a monetary and a …scal authority .4 1. In the …nal goods sector …rms produce di¤erentiated goods which are imperfect substitutes for goods produced abroad.4 1.6 Cyprus 0.0 1.9 Romania 0.0 2.5 0. A more detailed description of the model can be found in Roeger et al.6 2.4 1.0 2.6 Malta 0.8 2.2 2. Final good producers use a composite of domestic and imported intermediate 4 .0 1.7 Estonia 1.9 1. 3 Modelling the impact of Structural Funds The model used in this analysis is an extended version of the QUEST III R&D model. …nal and intermediate goods producing …rms.9 Czech Republic 1.Table 1: Payment Pro…le 2007-2015 Country 2007 2008 2009 2010 2011 2012 2013 2014 2015 Bulgaria 1.3 1.1 1.3 1.8 1.8 Source: own estimates.7 1.3 1.7 1.0 2.0 1.5 1.1 1.7 1.5 1. a research industry.3 1.4 Latvia 1.4 2.0 0.6 All NMS 1.2 0.1 The Model The model economy is populated by households.6 2.3 1.7 2.1 1.

Technological change is modelled as increasing product variety in the tradition of Dixit and Stiglitz (1977). Assis Human Capital goods and three types of labour . Households buy the patents of designs produced by the R&D sector and license them to the intermediate goods producing …rms. The production of new designs takes place in research labs. The model distinguises two types of households. Another share of households is liquidity-constrained. and they also buy the patents of designs produced by the R&D sector and license them to the intermediate goods producing …rms. employing high skilled labour and making use of the existing stock of domestic and foreign ideas.and high-skilled). These households cannot trade in …nancial and physical assets and consume their dispos- 5 .Figure 1: Shares of payments by …elds of interventions 100% 100% 80% 80% 60% 60% 40% 40% 20% 20% 0% BUL CZR EST LAT LIT HUN POL ROM SLO SLK NMS 0% Infrastructure RTD Services Industry Tech. The …rst group of households have access to …nancial markets where they can buy and sell domestic and foreign assets (government bonds). The intermediate sector is composed of monopolistically competitive …rms which produce intermediate products from rented capital input using the designs licensed from the household sector. These household members o¤er mediumand high-skilled labour services. accumulate physical capital which they rent out to the intermediate sector. medium.(low-.

the remaining structural parameters and variables are adopted from the available estimates in empirical studies (see Ratto et al. 2 6 . (2008)) or tied down by the equations of the model. For each skill group we assume that both types of households supply di¤erentiated labour services to unions which act as wage setters in monopolistically competitive labour markets. which constitute the key elements for modelling the Structural Funds interventions.t = KtP (1) The …nal good sector uses a labour aggregate (LY.t ) and intermediate goods (xi.2 3. we follow the literature of dynamic general equilibrium modelling and set the key steady-state ratios equal to their empirical counterparts for each region. where At X i=1 xi.1 Final goods production and public capital We account for the productivity-enhancing e¤ect of infrastructure investment via the following aggregate …nal goods production function: Yt = (1 At )( 1 1 ) 1 KtP LY. In calibrating the model. etc. One extension to the model made here is an explicit formulation of human capital accumulation following Jones (2002) in order to account for the signi…cant part of Structural Fund investments in various human resource programmes.t ) using a Cobb-Douglas technology.1. investment to output. While the calibration of the main steady state ratios (private consumption to output. Nominal rigidity in wage setting is introduced by assuming that households face adjustment costs for changing wages.t KtG G F CY . Our formulation assumes See the appendix for the description and the calibration of knowledge production and human capital accumulation equations. The unions pool wage income and distribute it in equal proportions among their members. Members of liquidity constrained households o¤er low-skilled labour services only. Below we describe in some more detail the supply side of the model and the …scal side.) is based on EUROSTAT and OECD data.able income each period. subject to a …xed cost F CY .

Intermediate goods producing …rms sell their products to domestic …nal good producers. It is t through this shock that we simulate the increase in infrastructure investment. therefore in a symmetric market framework the total output of the intermediate sector amounts to the total private capital stock as At P xi.2 Intermediate production and the R&D sector The intermediate sector consists of monopolistically competitive …rms which have entered the market by buying licenses for design from domestic households and by making an initial payment F CA to overcome administrative entry barriers.1.t = At xt = KtP . In symmetric equilibrium the inverse demand function of domestic …nal good producers 7 . One unit of intermediate goods is produced with an elasticity of substitution 1=(1 from one unit of private capital (KtP ). Capital inputs are also rented from the household sector for a rental rate of iK . Infrastructure investment is assumed to be proportional to output ItG = (IGSt + "IG )Yt t (3) where "IG is an exogenous shock to the share of government investment (IGSt ). i=1 Public infrastructure investment (ItG ) accumulates into the public capital stock K G according to KtG = (1 where G. Final output ( Yt ) is produced using At varieties of intermediate inputs ). Firms which have acquired a t design can transform each unit of capital into a single unit of an intermediate input.that investment in public capital stock (KtG ) increases total factor productivity with an exponent of G. G G )Kt 1 + ItG (2) the depreciation rate of public capital is set at 4 per cent. 3.

t xi.t = max pxi. Each domestic intermediate …rm solves the following pro…t-maximisation problem.is given as At X i=1 pxi. which is a prerequisite of production of innovative intermediate goods. The R&D sector hires high-skilled labour LA.t xi.t and generates new designs according to the following knowledge production function: 8 . the present discounted value of pro…ts is equated to the …xed entry costs plus the net value of patents 1 1 tK t 1 A 1 XY =0 j=0 PtA + A + F CA = 1 1 + rt+j x P Rt+ : (8a) For an intermediate producer. and the …xed entry cost F CA . x P Ri.t = where t t (1 )Y xj i.t t iA PtA F CA : (5) subject to a linear technology which allows to transform one unit of e¤ective capital (ki ucap) into one unit of an intermediate good x i = ki : (6) The no-arbitrage condition requires that entry into the intermediate goods producing sector takes place until x x P Ri.t 1 (4a) is the inverse gross mark-up of the …nal goods sector.t iK PtC ki.t ! 1 xi. entry costs consist of the licensing fee iA PtA for the design t or patent.t = P Rt = iA PtA + rt F CtA t (7) or equivalently. Innovation corresponds to the discovery of a new variety of producer durables that provides an alternative way of producing the …nal good.

3. The international stock of knowledge grows exogenously at rate gAw .t is composed of three skill-types of labour force: 9 .t (10) Therefore the …rst order condition implies: At LA.t : (9) In this framework we allow for international R&D spillovers following Bottazzi and Peri (2007).t+1 (11) where dt is the discount factor.t PtA = WtH + A WtH LA. i.At = At $1 At 1 LA. when innovation decreases with the level of knowledge. while can be interpreted as total factor measures the elasticity of R&D production on the number of researchers (LA ).3 Human capital accumulation The labour aggregate LY.e. Parameter e¢ ciency of R&D production. Negative value for these parameters can be interpreted as the "…shing out" e¤ect.t dt PtA At WtH LA. We assume that the R&D sector is operated by a research institute which employs high skilled labour at their market wage W H . Parameters $ and measure the foreign and domestic spillover e¤ects from the aggregate international and domestic stock of knowledge (A and A) respectively.t A 2 WtH L2 A. We also assume that the research institute faces an adjustment cost of hiring new employees and maximizes the following discounted pro…t-stream: 1 X t=0 max LA.1. while positive values refer to the "standing on shoulders" e¤ect and imply positive research spillovers. Note that = 1 would give back the strong scale e¤ect feature of fully endogenous growth models with respect to the domestic level of knowledge.t H dt Wt+1 LA.

The government provides subsidies (St ) Note that high-skilled labour in the …nal goods sector LHY is total high-skilled employment minus t the high-skilled labour working in the R&D sector ( LA.Y 1 L hH LHY t t L 1 L L L 1 : (12) Parameter ss is the population share of the labour-force in subgroup s (low-.LY. the parameter corresponds to the return to schooling estimated by Mincer (1974). 3. which we apply here. hs is the corresponding t accumulated human capital (e¢ ciency unit). mediumand high-skilled). The labour-market literature suggests that a reasonable value for is 0:07. 3 10 . Ls denotes the employment rate of population s. On the expenditure side we assume that government consumption (Gt ).2 The government budget constraint For the government sector various expenditure and revenue categories are separately modelled. The studied in a wealth of microeconomic research. >0 (13) The exponential formulation used here adapts Jones (2002) into a disaggregated skillstructure by incorporating human capital in a way that is consistent with the substantial growth accounting literature with adjustments for education4 . government transfers (T Rt ) and government investment (ItG ) are proportional to GDP and unemployment bene…ts (BENt ) are indexed to wages. Interpreting s t parameter has been as years of schooling. and L is the elasticity of substitution be- tween di¤erent labour types3 . Investments in human capital can then be modelled by increasing the years of schooling ( s t) for the respective skill-groups.t ). An individual’ human capital is produced by participating s in education and human capital : s t represents the amount of time an individual spends accumulating hs = hs e t s t . 4 See Barro and Sala-I-Martin (1995).t = sL 1 L hL LL t t L 1 L + sM 1 L hM LM t t L 1 L + sH.

Therefore. Total spending increases only by the amount of Structural Fund transfers. Additionality requires that Structural Funds are additional to domestically…nanced expenditure and are not used as a substitute for it. The co-…nancing principle means the EU provides only matching funds to individual projects that are part of the operational programmes and that the EU funds are matched to a certain extent by domestic expenditure. The additionality and co- 11 . the EU transfer COHt has to be matched by domestically-…nanced expenditure c:COH . Cohesion policy programmes are subject to the condition of additionality and co…nancing. government debt (Bt ) evolves according to and DEF are coe¢ cients. The problem with de…ning a proper benchmark means that in practice this principle of additionality is hard to verify and is thus not always binding. G Government revenues (Rt ) are made up of taxes on consumption as well as capital and labour income. Existing national resources that were used to …nance similar areas of interventions (and are thus concerned by the additionality requirement) can be ’ earmarked’ to co-…nance Structural Fund transfers. There is a lump-sum tax (TtLS ) used for controlling the debt to GDP ratio according to the following rule Bt Yt Bt Yt TtLS = B 1 1 bT B + DEF (14) where bT is the government debt target. Bt = (1 + rt ) Bt 1 + PtC Gt + T Rt + BENt + St G Rt TtLS COHt (15) It is assumed that the additional contributions to the EU budget are …nanced in the donor countries through an increase in lump-sum taxes. Fiscal transfers for NMS received from the EU are denoted by COHt (which is negative for the net contributors). Member States are not required to create new budgetary expenditure to co-…nance cohesion policy support.on physical capital and R&D investments in the form of a tax-credit and depreciation allowances. More formally. i.e. assume a co…nancing rate of c.

3. COHt is the …scal transfer received from the EU cohesion funds. In this case EXP0 > c COHt . c COHt ) (16) where T OT EXPt is total expenditure. 5 12 . EXP0 domestically– …nanced expenditure in the counterfactual situation (without Structural and Cohesion Funds). it is apparent that national public expenditure concerned by additionality usually exceeds the co-…nancing needs by far. this exercise takes domestically– …nanced expenditure EXP0 in the counterfactual situation (without structural and cohesion funds) as the benchmark and only examines the impact of the …scal transfer COHt received from the EU cohesion funds (equation 17). We assume that these shares of the …elds of interventions are constant for all the years of the payment horizon 2007-2015. Examining the additionality tables of Member States. Herve and Holzmann (1998) criticise earlier model-based studies of structural funds for grossly exaggerating the total impact because they assumed that the full Structural Fund spending is additional to investment in the counterfactual situation T OT EXPt = COHt + c COHt + EXP0 while the correct formulation of the additionality principle is given by equation (17). and c is the co-…nancing rate. and total expenditure is given by5 T OT EXPt = COHt + EXP0 (17) As spending on infrastructure and education is already high in the NMS countries. The transfers are modelled as lump-sum transfers between governments.3 Implementing the interventions The …scal transfers under the Structural and Cohesion Policy programmes amount to additional injections of up to 3 percent of GDP for the NMS.…nancing principles can be expressed as the following condition for total government spending in a bene…ciary country: T OT EXPt = COHt + max(EXP0 . their respective share in the total amount of interventions and the way each of the interventions are captured as shocks to the model. Table 2 shows the main …elds of interventions.

On the basis of available data on current spending on education (around 5 per cent on all skill-groups and 1 per cent on high-skilled in terms 13 .Table 2: Fields and Shares of Interventions Share Variable to implement the shock 62:0 Temporary increase in I G . government investment (via "IG ) t Industry&Services 9:1 Reducing …xed costs faced by …nal goods …rms (F CY ) 1:5 permanent reduction F CY (categories 13.14. via decreasing the …xed costs faced by the intermediate sectors (F CA ) temporarily or permanently. Concerning human capital investments we distinguish three subcategories of payments based on the detailed payment pro…le. R&D promoting spending is modelled similarly. Investment in public infrastructure is modelled via a temporary increase in government investments "IG . 15 of the payment pro…le) 7:6 temporary reduction F CY RTD 10:1 Reducing the …xed costs faced by the users of R&D products (F CA) 4:5 permanent investment in R&D infrastructure (categories 2 and 3 of the payment pro…le) 5:6 temporary investments Human resources 15:3 Raising human capital and government consumption expenditures 1:4 investment in high-skilled human capital (hH via H ) t t (category 74 of the payment pro…le) 4:8 educational investments in all skills (hs via s ) t t (categories 72 and 73 of the payment pro…le) 9:1 other government expenditures (Gt ) Technical assistance 3:4 Temporary increase in government consumption (Gt ) Field Infrastructure Source: European Commission (DG REGIO). and allocated in the model to all skill groups.4 per cent is directly targeting investments in high-skilled human capital and captured in the model as a shock to H t : The remainder is accounted for as temporary increase in government consumption (Gt ). depending on the nature of the programme.8] per cent of the funds are spent on educational investments without speci…c skill-speci…cation. Support to industry and services-related programmes are introduced t via a temporary or (depending on the nature of the programme) permanent decrease in …xed costs of …nal goods …rms (F CY ). The detailed categories of the payment pro…le are shown in the Annex. A smaller share of 1. Around [4.

as becomes 6 See the Appendix for a detailed description of the calibration of human capital accumulation.of GDP) an estimate can be made of the additional years of schooling (increment to s t) that can be …nanced by the …scal transfers6 . or ideas. in particular in government investment. But the output gains become gradually larger as the supply side e¤ects become stronger. Higher demand for patents (ideas) increases the demand for high skilled workers and boost innovation. with private investment rising proportionally less than GDP. It takes years for these supply side e¤ects to build up and in the …rst few years the demand e¤ects dominate. As a result of this. and a gradual accumulation of public capital. By 2020. These positive supply side e¤ects are permanent. There is a large increase in government spending. the stock of "knowledge". R&D promoting policies lower entry costs and so reduce the pro…t requirements for intermediate producers and encourages entry of new …rms. in the …rst two years GDP rises by less than the size of the demand impulse. Investment in training and education boost human capital and raises output further. and by 2015 output rises to more than 4 per cent above baseline. In order to account for the additional time spent on training. the additional spending puts upward pressure on wages and interest rates and leads to some crowding-out. we assume that the last cohort of student population stays longer in the education system and enter into the active labour force later. Finally technical assistance is introduced as a temporary increase in government consumption. This raises productivity and leads to a gradual increase in output in the …nal goods sector. is up by 2 per cent. Growth is further boosted by the other interventions. 14 . 4 4. In the short run. The incentives shift employment of high skilled workers from the …nal goods sector to the R&D sector.1 The potential macro economic impact Simulations results The interventions of the Structural and Cohesion Policy programmes are simulated in the model as shocks described above and Table 3 shows the macroeconomic impact on the new member states.

7 Consumption of non-liquidity-constrained households is directly boosted by higher expected future income and aggregate consumption (liquidity and non-liquidity constrained) is already in the …rst year more than 1 per cent higher. As this e¤ect is ignored in model simulations. 15 . it could lead to an increase in the participation rate. This leads to some relocation of high skilled workers from production of …nal goods to the R&D sector and an increase in the wage of high skilled workers. this subsidy has only a small e¤ect in the …rst years of the programmes because of crowding-out due to overproportionally higher government spending on infrastructure and human resources which puts upward pressure on real interest rates. a signi…cant share of the demand impulse leaks abroad through higher imports. but output is permanently higher. the GDP e¤ect for the donor countries is negative. Finally. Imports are more than 2 per cent higher and the trade balance deteriorates. In the long run the total employment e¤ect remains negative because of a negative terms of trade e¤ect. we potentially underestimate the impact on employment rates. but the relatively smaller output loss indicates a small positive spill-over e¤ect from higher growth in the new member states. Although the business sector receives direct support from the government. when the programmes are assumed to be terminated and there is no longer a demand impulse.apparent when looking at the e¤ects after 2015. but the employment e¤ect is slightly negative afterwards. In small open economies. 7 To the extent that human capital investment improves the employability of the non-active population. The demand impulse leads to a small appreciation in the short run but as the productivity e¤ects become stronger the e¤ective exchange rate depreciates.1 per cent of GDP. R&D enhancing policies account for signi…cant increase in the employment share of high-skilled labour in the research sector. Employment rises in the …rst years due to the demand expansion. due to an increase in their contributions to the EU budget of around 0.

21 -1.01 -0.14 0.31 -0. .55 -1.37 0.06 -0.37 -2.24 0.38 3.07 0.23 2009 0.82 0.00 0.03 -0.90 -0.88 0.11 0.01 -0.07 -0.medium .01 0.42 2.21 -0.84 2.90 -2.17 -0.03 0.42 2.25 0.20 -1.58 2014 3.20 -0.30 -2.91 -1.03 0.00 2.14 0.19 2.77 0.13 -0.04 0.99 -0.22 -0.92 -0.00 0.85 2.34 -0.05 0.28 0.04 0.49 2.13 -0.77 1.94 2.26 2.31 -0.price level Terms of trade REER Euro exchange rate Nom.17 1.03 1.89 -0.44 -0.20 -1.19 -0.35 2.03 0.04 -0.01 0.07 0.39 2.03 0.54 -2.48 -1.02 -0.58 -0.03 -0.28 3. .03 3.01 0.38 -1.30 3.59 0.01 0.33 -0.02 0.17 2.01 -0.01 2. interest rate Real interest rate In‡ ation Cons.55 -0.04 Note: Percentage (top half) and absolute (bottom half) di¤erences from baseline 16 .00 -0.51 -0.05 5.89 -0.58 -0. productivity Employment rate .32 -0.10 0.32 2.81 2.79 0.18 0.02 0.06 -0.12 0.04 0.55 0.03 1.04 0.32 -0.42 0.45 -0.41 2011 1.29 2.97 0.16 -0.19 0.92 -0.21 0.40 -0.21 1.16 -0.61 -2.65 0.43 -0.04 1.36 0.48 0.03 0.40 0.60 2.01 0.61 -0.01 -0.41 -0.12 -1.42 -0.05 0.35 -0.08 -0.41 0.09 0.03 4.debt (%GDP) Gov.04 -0.02 0.(…nal goods) .23 0.28 -0.03 -0.12 -0. .29 -0.68 1.65 -0.08 -0.04 0. payments.55 -0.08 0.65 0.01 -0.35 -0.02 -0.87 0.11 -0.26 -1.29 1.07 2.23 0.01 0.42 -0.11 0.36 -3.02 0.02 -0.40 1.05 0.06 -0.06 -0.65 -0.34 -0.26 -0.07 2.52 0.71 -0.13 -0.20 -0.01 2.15 -0.49 1.00 0.26 -0.07 Euro-area GDP 0. .06 0.56 -0.38 0.64 0. Investment Exports Imports Real wages Price level Consumer.12 0.32 0.17 0.25 2.17 0.47 1.09 1.15 -0.31 -0.low.00 -0.01 0.91 -0.42 -0.86 -0.01 3.03 -0.06 -0.07 6.00 0.non.07 -0. hh.08 3.55 3.09 -0.83 -0.20 0.54 0.06 0.06 -0.28 -0.24 -0.22 -0.50 -1.32 0.01 -0.02 0.constr.80 -1.63 1.73 -0.19 -0.01 0.33 1.13 0.49 1.03 -0.19 -1.Table 3: Simulated macroeconomic e¤ects Variable GDP Ideas TFP Human cap.01 -0.47 -0.04 -0.01 2.05 -0.33 3.47 -0.16 -0.23 1.25 2.01 0.08 -1.74 0.44 -1.18 -1.02 -0.05 0.09 0.05 2016 3.02 -0.25 2010 1.00 -0.44 0.32 0.high (…nal goods) .low Human cap.77 3.41 3.03 1.01 -1.45 -0.16 -1.99 3.06 0.16 2.02 1.77 0.98 0.47 1.08 -0.12 -0.00 0.32 -0.01 1.23 0.20 -1. in‡ ation R&D intensity Lab.high Coh.55 2.29 -1.medium (…nal goods) .32 1.58 3.39 -0.08 -0.97 2.52 -0.low .24 0.31 1.45 -0.09 -0.00 0.09 -0.93 3.09 2020 3.high Employment .75 -0.30 -0.32 0.medium Human cap.44 3.19 1.20 -0.44 -0.13 -0.09 0.08 0.19 -0.25 1.23 -0.59 0.02 -0.00 -1.00 1.14 0.06 6.02 -0.00 -0.02 0.10 0.44 3.37 -0.06 -0. (%GDP) Gov.05 -0.25 -0.04 0.16 -0.12 -0.23 0.05 -0.04 5.14 -0.38 3.09 2.02 1.01 0.70 -1.47 2012 2.54 -3.04 -0.27 0.02 1.12 0.79 -0.09 -0.80 0.00 2.27 0.98 -1.06 0.88 2.61 -0.01 -0.20 2.23 -0.39 -0.05 -0.31 0.04 -0.67 1.78 -1.12 -0.hh.23 -0.00 -0.80 0.20 1.03 -0.28 1.73 0.42 1.34 0.60 -0.06 -0.43 -0.11 -0.54 -0.constr.62 2015 3.27 1.high (R&D sector) Consumption .44 -0.53 0.34 -0.99 0.48 1.44 -0.13 -0.40 0.balance (%GDP) Trade balance (%GDP) 2007 0.57 -0.30 0.04 1.52 -0.43 0.04 -0.06 -1.07 -0.35 1.liq.05 0.30 1.96 2008 0.35 3.30 -0.97 0.78 -2.65 0.08 -0.09 -0.18 -2.96 -3.56 2013 2.73 -0.37 0.66 2.04 -0.45 3.64 -0.02 0.98 -1.09 -0.58 2.89 -0.

The benchmark assumption in the model is that the rate of return on public capital equals that of private capital (Gramlich (1994). simultaneity bias and reverse causation. Total interventions classi…ed as human capital investment in the Cohesion and Structural Funds account for 15 per cent of total spending but this category covers a wide range of measures. various measures improving the social inclusion of less-favoured groups. The projected output gain in 2015 can be as high as 5 per cent under a more favourable assumption of when G G = 0:15. see e.2 Sensitivity analysis The results are sensitive to the assumption on the marginal product of public capital. There is much uncertainty about the productive impact of infrastructure investment and econometric studies show large variation in estimates. of around a quarter of the benchmark e¤ect. p. Many of the early empirical estimates have been dismissed as implausibly high 8 . depending how care is taken of common trends.4. It is not a priori clear whether all these measures improve skill e¢ ciencies as much as assumed in the model simulations.(see annex). it can take from 1 to 6 years till the supply side improvements raise the level of GDP above the level of the direct demand injection from the programmes. The results are also highly sensitive to how we interpret the interventions on human capital formation. environmental protection and development of cultural and social infrastructure) it could be argued that this assumption is too optimistic.g. Depending on the magnitude of the marginal productivity of capital. However.g. A sensitivity analysis to the speci…c assumptions concerning this category shows the impact these assumptions have on the overall results.g.1187). the surveys by Gramlich (1994). including e. 8 17 . for some of the infrastructure interventions (e. missing variables. The sensitivity analysis shows substantial variations both in the peak as well as in the long-run GDP e¤ects. Our benchmark scenario assumes that only those categories that are explicitly labelled as improving human capital in the programmes For an overview of the literature and the econometric problems. Figure 2 plots the GDP e¤ect of the programmes under three alternative scenarios relative to our baseline assumption for G (continuous middle line in the …gure). while it could be as low as 3 per cent = 0:05.

e. benchmark alphag=0. 6.10.05 alphag=0. Our alternative scenario shown in Figure 2 plots the GDP e¤ect if all spending on human resources (15. 18 . categories 72.Figure 2: GDP e¤ects under di¤erent assumptions on the marginal productivity of public capital (% di¤erence from baseline) 7 7 6 6 5 5 4 % 3 4 3 2 2 1 1 0 20 07 Q 20 1 07 Q 20 3 08 Q 20 1 08 Q 20 3 09 Q 20 1 09 Q 20 3 10 Q 20 1 10 Q 20 3 11 Q 20 1 11 Q 20 3 12 Q 20 1 12 Q 20 3 13 Q 20 1 13 Q 20 3 14 Q 20 1 14 Q 20 3 15 Q 20 1 15 Q 20 3 16 Q 20 1 16 Q 20 3 17 Q 20 1 17 Q 20 3 18 Q 20 1 18 Q 20 3 19 Q 20 1 19 Q 20 3 20 Q 20 1 20 Q3 0 quarters Net-transfers received (% GDP) alphag=0. it gives a clear upper bound of the potential impact this spending can have on output.73 and 74.15 (i. While this assumption may be too optimistic given the social nature of some of this spending.2 per cent of total spending) are a¤ecting human capital accumulation in the model.3 per cent of toatl spending) is assumed to be educational investments to improve human capital at each skill-level.

and in that sense our results give an 19 .Figure 3: GDP e¤ects under more optimistic scenario on human capital investments (% di¤erence from baseline) 7 7 6 6 5 5 4 % 3 4 3 2 2 1 1 0 20 07 Q 20 1 07 Q 20 3 08 Q 20 1 08 Q 20 3 09 Q 20 1 09 Q 20 3 10 Q 20 1 10 Q 20 3 11 Q 20 1 11 Q 20 3 12 Q 20 1 12 Q 20 3 13 Q 20 1 13 Q 20 3 14 Q 20 1 14 Q 20 3 15 Q 20 1 15 Q 20 3 16 Q 20 1 16 Q 20 3 17 Q 20 1 17 Q 20 3 18 Q 20 1 18 Q 20 3 19 Q 20 1 19 Q 20 3 20 Q 20 1 20 Q3 0 quarters Net-transfers received (% GDP) Human capital. upper-bound 5 Conclusions This paper descibes how a DSGE model with endogenous growth can be used for an ex-ante (prospective) evaluation of the potential impact of EU Cohesions and Structural Funds. There are strong reasons to expect at least part of the total spending to be diverted to sub-optimal use (Herve and Holzmann (1998)). An important caveat for model based analyses like these is that they only tell us something about the potential impact of EU Funds. benchmark Human capital. And the model applies the Jones (1995) semiendogenous growth framework to explicitly model the underlying development of R&D. which allows us to analyse the impact of R&D promoting policies on growth. The model simulates the impact of increased public infrastructure investment and captures the productivity enhancing e¤ects of this. The model incorporates endogenous human capital accumulation and simulates the e¤ects of policies promoting vocational education and training on skill e¢ ciencies. assuming an e¢ cient and optimal use.

Ederveen. J. The international dynamics of R&D and innovation in the long run and in the short run. Infrastructure investment: A review essay. 67(3):297– 308.pdf. (2002). and Nahuis. and Shleifer. 117(1):1– 37. Ederveen. Lopez-De-Silanes. de Mooji. F. and Peri. Promotion of tourism. and Sala-I-Martin. American Economic Review. 117(3):486– 511. Blackwell Publishing. Gramlich. A senstitivity analysis shows how our results depend on assumptions related to the productivity of various forms of infrastructure investment and that of policies promoting vocational education and training. Journal of Economic Literature. S.. Bradley. Analysis of the impact of cohesion policy. 17– 42. A. Much uncertainty exists on these productivity e¤ects and more micro analysis is clearly needed to shed light on these e¤ects.. Bottazzi. CPB Netherlands’Bureau for Economic Policy Analysis. W. L. Djankov. X. It should also be recognised that the objective of EU Cohesion Policy is economic and social cohesion. The Economics Journal. Funds and games: The economics of european cohesion policy.. and Mitze. E.europa. The regulation of entry. and Nahuis. T.. 20 ..upper bound of the likely e¤ects. R. and hence wider than economic growth per se. http://ec. A. (1994). Porta.. culture. 32(3):1176– 1196. R. Untiedt. (2002). Monopolistic competition and optimum product diversity. R. (2007). Fertile soil for structural funds? a panel data analysis of the conditional e¤ectiveness of european cohesion policy. S. (1995). and Coleman. and Stiglitz. The Hague. Economic Growth. (2006). Caselli.. J. (2006). Technical note.. McGraw Hill. Kyklos.eu/regionalpolicy/sources/docgener/evaluation/pdf/hermin07. F. J. American Economic Review. R. K.. References Barro. R. E. G. G. a note explaining the HERMIN-based simulations. Dixit. Groot. H. Future research should also focus on how policies promoting social inclusion increase participation in the labour force and how these e¤ects can be incorporated in the model. The Quarterly Journal of Economics. J. Special Publication 41. J. S. (1977). Groter. urban and rural regeneration are equally important policy objectives. The world technology frontier. New York. 96(3):499– 522. 59(1):pp.. (2007). L.

Y. C. Brussels. Center for European Policy Studies. (2008). J. 92(1):220– 239. (2002). M. Schooling. (1992). J. C. economic growth in a world of ideas. Portuguese Economic Journal. Jones. Pessoa. Fiscal Transfers and Economic Convergence in the EU : An Analysis of Absorption Problems and an Evaluation of the Literature. Columbia University Press. Jones. Paris. Herve. Ideas driven growth: The OECD evidence. I. (1998). Katz. OECD Science. American Economic Review. How to achieve a better budget for the european union? CEPS Working Document 289. (1995). Journal of Political Economy. OECD (2006). (also published as: European Economy Economic Paper no. A. Nomos Verlagsgesellschaft. QUEST III: An estimated DSGE model t of the euro area with …scal and monetary policy. J. L. and Earnings. J.S. Quarterly Journal of Economics. (2008). 4(1):46– 67.. Roeger. Technology and Industry Working Papers 2006/4. Experience. M.. Brussels.. and Murphy. D. (1974). Source of U. Baden-Baden. Economic Modelling (forthcoming). (2006). R&D-based models of economic growth. and in ’ Veld. K. R. July 2008). 103(4):759– 84. Mincer. 21 .Gros. I. Ratto. F. and Holzmann. Warda. Structural reforms in the EU: A t simulation-based analysis using the QUEST model with endogenous growth. European Commission Directorate-General for Economic and Financial A¤airs. Tax treatment of business investments in intellectual assets: An international comparison. Changes in relative wages. OECD. 1963-1987: Supply and demand factors. Varga. Roeger. J. 107(1):35– 78. (2008). W. (2005). and in ’ Veld. OECD. Education at Glance.335 . Paris.. W. European Economy Economic Paper 351. New York. European Commission Directorate-General for Economic and Financial A¤airs. Brussels.

Innovation corresponds to the discovery of a new variety of producer durables that provides an alternative way of producing the …nal good. Parameters $ and 22 . a monetary and a …scal authority. a research industry. …nal and intermediate goods producing …rms. This form of R&D equation accounts for international spillovers almost identically to the speci…cation of Bottazzi and Peri (2007).t At PY Yt 1 A (d) iA. The driving equation system of the semi-endogenous technological change can be summarized as At = At $1 At 1 LA.Appendix: Calibrating the parameters of knowledge production. The production of new designs takes place in research labs. medium-.t (a) 1 + gA = (1 + gn ) 1 (b) PA. The R&D sector hires high-skilled labour (LA ) and makes use of the existing stock of domestic and foreign ideas to generate new designs ( At ). The model economy is populated by households. The intermediate sector consists of monopolistically competitive …rms which have entered the market by licensing a design from domestic households (at rate of iA. Capital inputs are also rented from the household sector for a rental rate of iK . employing high skilled labour and making use of the existing stock of ideas. Final good producers use a composite of intermediate goods and three types of labour . intermediate goods production and human capital accumulation Our semi-endogenous model builds on Jones (1995) version of R&D modelling.t At = wH LA.t ) and by making an initial payment F CA to overcome administrative entry barriers. Firms which have acquired a design can transform each unit of t capital into a single unit of an intermediate input.t (c) rdi = PA. but we account for international R&D spillovers following Bottazzi and Peri (2007). and high-skilled).(low-.t + rt F CA = iA = (1 A t. Households buy the patents of designs produced by the R&D sector and license them to the intermediate goods producing …rms. where A t+1 + tK ) A t = 1 xt (e) )(it (1 ) tK + rpA t (f) (g) K t = At x t The …rst equation is the spillover-augmented version of Jones (1995) R&D production.t PA.

t ) and on the R&D intensity A. equation (c) shows the …rst order condition of R&D production. Decreasing entry costs lowers the pro…ts requirement for intermediate producers and thus increases entry of new …rms and the demand for patents.measure the foreign and domestic spillover e¤ects from the aggregate international and domestic stock of knowledge (A and A) respectively. We assume that the R&D sector is operated by a research institute which employs high skilled labour at their market wage wH . equation (g) states the identity between the total intermediate goods production and physical capital under symmetric equilibrium. A. P the values of gA and gn are given in our baseline model9 . 1. Equation (b) states the balanced-growth relationship between the growth of ideas gA (= gAw ) and population gn . and $: = and = $long 1 term long term 1 long term : long term The estimated values for these two coe¢ cients show fairly big variations under the di¤erent regressions. The population growth gn is obtained from EUKLEMS potential output calculations. The authors do not estimate directly and $. equation (d) is the de…nition of R&D-intensity: total R&D expenditure of the intermediate sector in percentage of GDP. The international stock of knowledge grows exogenously at rate gAw . 9 23 . and it might be inadequate to apply these long-term coe¢ cients on our Pessoa (2005) provides estimates for the growth of patents or ideas in various OECD countries at an average of gA = 0:057. when innovation decreases with the level of knowledge. and . satisfying the following theoretical restrictions between the long-term coe¢ cients of . $. Equation (e) states the freeentry condition between the pro…t of the intermediate sector ( t ). Negative value for these parameters can be interpreted as the "…shing out" e¤ect. R&D production Although we do not have direct estimates of . we can use the existing literature and the model restrictions to get calibrated values for them. while measures the elasticity of R&D production on the number of researchers. In the second step we use the estimated long-term relationship between and from Bottazzi and Peri (2007) to approximate $ separately. Since one unit of capital is used to produce one unit of intermediate good (xt ). Data on the R&D P AD share of labour ( LA. Note that = 1 would give back the strong scale e¤ect feature of fully endogenous growth models with respect to the domestic level of knowledge. while positive values refer to the "standing on shoulders" e¤ect and imply positive research spillovers.e. and the per unit price of R&D inventions (PA ) and the …xed (entry) cost F CA . In order to set and $ in the …rst step we express the sum of these two parameters from equation (b). Equation (f) de…nes the rental rate of intangible capital which takes into account that households pay income tax at rate tK on the period return of t intangibles and they receive tax subsidies at rate A .tY Yt t is obtained from EUROSTAT. i. however their estimated cointegration vector contains two coe¢ cients and . Parameter can be interpreted as total factor e¢ ciency of R&D production. These values together with the restrictions of the balanced growth dynamics and the other variables of the baseline pin down and PA .

Ls denotes the employment rate of population s.t = sL hL LL t t L 1 L + sM hM LM t t L 1 L + sH. t A. where the subscripts denote the skill-groups (low-L. However the ratios of these two coe¢ cients = $long term long term vary less. The authors argue in The full sample consists of …fteen OECD countries including the US and ten member states of the European Union. All other variables given. The B-indices published in Warda (2006) can be t applied to calibrate A and tK . 11 We use the net mark-up of the manufacturing sector calculated in EUKLEMS to obtain . medium-M and high-H). In the last step we subtract this value from the sum of and $ as we calculated from equation (b) earlier.and high-skilled labour. Intermediate goods production The calibration of the parameters in intermediate goods production relies on the entry costs estimations of Djankov et al. the arbitrage equation (e) determines the rental rate of intangible capital. The pro…t of a representative intermediate …rm is determined by its production and the net mark-up of the sector11 . 2. Finally.t ). medium. imposing the ratio of the long-term parameters instead of their exact values is also less restrictive. rpA . Finally. ss is the population share of labour-force in subgroup s.Y hH LHY t t L 1 L ."contemporary" speci…cation. 10 24 . A. we use the ratio of these parameters from the speci…cation in which the authors omitted the US from their regressions10 . and the estimations for R&D related subsidies ( A ) of Warda (2006). and L is the elasticity of t substitution between di¤erent labour types. Given that we normalized the stock of domestic ideas to one (At ). We follow Caselli and Coleman (2006) which analyzed the cross-country di¤erences of the aggregate production function when skilled and unskilled labour are imperfect substitutes. (2002). hs is the skill-speci…c e¢ ciency unit of labour. 3. The elasticity of substitution between di¤erent labour types ( L ) is one of the major issue addressed in the labour-economics literature. we use the de…nition of equation (f) to obtain as residual the calibrated approximation of the risk-premium on intangibles. the inverse of the gross mark-up in the intermediate sector. Data on skill-speci…c population shares. The calibration is mostly based on EUROSTAT and OECD data. we normalize the stock of domestic and foreign ideas to one and therefore the values for and can be obtained from expressions (a) and (e). furthermore. Note that high-skilled labour in the …nal goods sector is the total high-skill employment minus the high-skilled labour working for the R&D sector (LA. Human capital accumulation Labour force is disaggregated into three skill-groups: low-. participation rates and wage-premiums are obtained from the Labour Force Survey and Science and Technology databases of EUROSTAT. iA . equation (g) pins down the per …rm quantity of intermediate goods production. The CES-aggregate for labour have the following form: 1 L 1 L 1 L L L 1 LY. To approximate our $ for the EU27.

in the baseline we set the variables of training s. s for the respective skill-groups are obtained from OECD t (2006). labour-types and e¢ ciency units: wM wL L L 1 hM t = sM LM sL LL 1 L 1 hL .T + lt R . The evolution of school population can be written as N Ss (t) = (1 and for the working population N Ws (t) = (1 s ) N Ws (t s + bs ) N Ss (t 1).T R s ) lt 1 + "s.: s t = s s. bs is the birth-rate of skill group s. s 2 fL. 4. For simulation purposes.e. s . Combining the dynamic 25 . then s = 1=80. The exit rate is the inverse of average duration spent in a skill-group. In order to simulate the educational investments in human capital we increase the years of schooling ( s ) for the respective skill-groups by t the additional years of schooling that can be …nanced from the …scal transfers (shock to "s. in year t. Finally. and s determine the evolution of the (inverse) dependency ratios IDEP RAT Es (t) = N Ws (t) N Ss (t) and the growth rate of working age population (gpopws (t)). and s is the exit rate from working age population. Hg respectively the number of students and workers in skill-group s. s is the exit-rate of the working age population. and "s. the participation in trainings can be interpreted as an addition to the years of schooling with a depreciation according to the exit rate of working age population. t A.T R . where s is the exit rate from student skill cohort s. we …x the return to schooling parameter of at 0:07.T where lt R = (1 s.T R to zero and given the years of schooling from OECD (2006) we can compute t hs from the de…nition of e¢ ciency. The parameters bs . 1) + s N Ss (t 1). In line with Jones (2002). t and hH t = wH wM L L 1 sH LH sM LM 1 L 1 hM : t In the next step we adapt Jones (2002) into a disaggregated skill-structure and impose s that the functional form of hs = hs e t describes the evolution of skill-speci…c human t capital. t where for each skill-group s.favour of using the Katz and Murphy (1992) estimate of 1:4. s. M. We normalize the e¢ ciency of low-skilled at 1 the other e¢ ciency units are restricted by the labour demand equations which imply the following relationship between wages. The number of school years.T regular education system. lt R is the year equivalent of the average time spent in training in period t. Population dynamics Denote by N Ss (t) and N Ws (t). i. s is the average number of years of schooling in the s.T lt R and "s.T R is the average yeart equivalent of training in period t.T R ). if the duration of education for high skilled is 20 years (80 quarters). For example.

For the calibration. therefore the growth rate of the working age population is given by N Ws (t) s = N Ws (t 1) IDEP RAT Es (t Finally.M. by adding the de…nitions shs (t) = P N Ws (t) s2fL. 26 .Hg N Ws (t) 1) s: we obtain the evolution of the skill population shares.equations of the student and working age population we get IDEP RAT Es (t) = 1 1 s s + bs IDEP RAT Es (t 1) + s 1 s + bs . we use the education statistics of OECD (2006) and data on the skill-distribution of working age population from EUROSTAT.

e-government.55 0. Fields of interventions.78 1. networking. existing R&TD centres and …rms. Assistance to R&TD. etc. establishment of new …rms by universities. e-inclusion. innovation and entrepreneurship R&TD activities in research centres R&TD infrastructure and centres of competence in a speci…c technology Technology transfer and improvement of cooperation networks between small businesses (SMEs). innovation.13 10 11 INF INF 0. research.08 1. etc.) Other measures for improving access to and e¢ cient use of ICT by SMEs % of total 01 02 03 RTD RTD RTD 1. integration of clean technologies into …rm production Investment in …rms directly linked to research and innovation (innovative technologies. e-content. etc).88 1.17 0.73 1. adoption and use of pollution prevention technologies. security. between these and other businesses and universities. detailed tables Priority themes Research and technological development (R&TD).) Other investment in …rms Other measures to stimulate research and innovation and entrepreneurship in SMEs Information society Telephone infrastructures (including broadband networks) Information and communication technologies (access.65 0. education and training. interoperability.21 2. research centres and scienti…c and technological poles (scienti…c and technological parks.) Information and communication technologies (TEN-ICT) Services and applications for citizens (e-health.Annex: Detailed payment pro…le categories Code Category Table A1. e-learning.45 27 .32 04 05 06 RTD S I 0. postsecondary education establishments of all kinds regional authorities.46 07 RTD 2.85 08 09 I RTD 2. etc. risk-prevention. etc. particularly in SMEs (including access to R&TD services in research centres) Advanced support services for …rms and groups of …rms Assistance to SMEs for the promotion of environmentally-friendly products and production processes (introduction of e¤ective environment managing system.75 1.) Services and applications for SMEs (e-commerce. technopoles.05 12 13 14 15 INF S S S 0.

19 0.94 6.00 0.30 0.28 28 .15 0.58 0.22 0.37 3.25 1.29 0.25 0.27 0.54 0. geothermal and other Energy e¢ ciency. co-generation.87 8.08 0.40 1.43 0.06 0.69 0.Code 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 Category INF INF INF INF INF INF INF INF INF INF INF INF INF INF INF INF INF I INF I INF I INF I I I I I Priority themes Transport Railways Railways (TEN-T) Mobile rail assets Mobile rail assets (TEN-T) Motorways Motorways (TEN-T) National roads Regional/local roads Cycle tracks Urban transport Multimodal transport Multimodal transport (TEN-T) Intelligent transport systems Airports Ports Inland waterways (regional and local) Inland waterways (TEN-T) Energy Electricity Electricity (TEN-E) Natural gas Natural gas (TEN-E) Petroleum products Petroleum products (TEN-E) Renewable energy: wind Renewable energy: solar Renewable energy: biomass Renewable energy: hydroelectric.52 0.18 0.96 0.47 3.09 0. energy management % of total 0.23 0.14 0.78 0.

45 0.57 54 55 56 57 58 59 60 61 INF S INF S INF INF S INF 0.35 1.23 0.31 0.17 0.52 29 .77 2.82 0.78 0.60 2.56 1.11 2.Code 44 45 46 47 48 49 50 51 52 53 Category INF INF INF INF INF INF INF INF INF INF Priority themes Environmental protection and risk prevention Management of household and industrial waste Management and distribution of water (drink water) Water treatment (waste water) Air quality Integrated prevention and pollution control Mitigation and adaption to climate change Rehabilitation of industrial sites and contaminated land Promotion of biodiversity and nature protection (including Natura 2000) Promotion of clean urban transport Risk prevention (including the drafting and implementation of plans and measures to prevent and manage natural and technological risks) Other measures to preserve the environment and prevent risks Tourism Promotion of natural assets Protection and development of natural heritage Other assistance to improve tourist services Culture Protection and preservation of the cultural heritage Development of cultural infrastructure Other assistance to improve cultural services Urban and rural regeneration Integrated projects for urban and rural regeneration % of total 2.01 1.37 0.48 0.13 0.52 5.

43 0. including through action to achieve a reduction in early school leaving. such as facilitating access to childcare and care for dependent persons Speci…c action to increase migrants’participation in employment and thereby strengthen their social integration Improving the social inclusion of less-favoured persons Pathways to integration and re-entry into employment for disadvantaged people. and development of systems for anticipating economic changes and future requirements in terms of jobs and skills Improving access to employment and sustainability Modernisation and strengthening labour market institutions Implementing active and preventive measures on the labour market Measures encouraging active ageing and prolonging working lives Support for self-employment and business start-up Measures to improve access to employment and increase sustainable participation and progress of women in employment to reduce gender-based segregation in the labour market. enterprises and entrepreneurs Development of life-long learning systems and strategies in …rms. gender-based segregation of subjects and increased access to and quality of initial vocational and tertiary education and training Developing human potential in the …eld of research and innovation.37 70 HC 0.25 0.36 30 .86 73 HC 1.56 63 64 HC HC 0. introduction and implementation of reforms in education and training systems in order to develop employability. training and support in connection with restructuring of sectors and …rms.Code Category Priority themes Increasing the adaptability of workers and …rms.75 65 66 67 68 69 HC HC HC HC HC 0. and networking activities between universities.50 0. in particular through post-graduate studies and training of researchers. combating discrimination in accessing and progressing in the labour market and promoting acceptance of diversity at the workplace Improving human capital Design.50 72 HC 2. research centres and businesses % of total 62 HC 1. improving the labour market relevance of initial and vocational education and training. and to reconcile work and private life.03 71 HC 1.57 1.76 0. training and services for employees to step up their adaptability to change.95 74 HC 1. promoting entrepreneurship and innovation Design and dissemination of innovative and more productive ways of organising work Development of speci…c services for employment. updating skills of training personnel with a view to innovation and a knowledge based economy Measures to increase participation in education and training throughout the life-cycle.

monitoring and inspection Evaluation and studies.00 85 86 2.00 0.71 0.32 2.24 82 83 84 S S I 0.Code 75 76 77 78 79 Category INF INF INF INF INF Priority themes Investment in social infrastructure Education infrastructure Health infrastructure Childcare infrastructure Housing infrastructure Other social infrastructure Mobilisation for reforms in the …elds of employment and inclusion Promoting the partnerships.00 0. implementation.39 0.19 0. regional and local level Mechanism for improving good policy and programme design. pacts and initiatives through the networking of relevant stakeholders Strengthening institutional capacity at national. information and communication % of total 2.19 0.74 31 .85 80 HC 0.16 81 HC 1. monitoring and evaluation Reduction of additional costs hindering the outermost regions development Compensation of any additional costs due to accessibility de…cit and territorial fragmentation Speci…c action addressed to compensate additional costs due to size market factors Support to compensate additional costs due to climate conditions and relief di¢ culties Technical assistance Preparation.