PUBLIC INFRASTRUCTURE: DEFINITION,

CLASSIFICATION AND MEASUREMENT ISSUES

GIANPIERO TORRISI
Newcastle University
[email protected]

ABSTRACT. Beginning from the end of the 1980s, many studies analysing
the relationship between infrastructure endowment and economic develop-
ment have been realised. A general consensus is achieved around the idea
that basic infrastructure facilities are important features related to economic
performance, although both magnitude and causality direction are debated.
A peculiar feature of these studies is that, across them, different empirical
and theoretical entities are referred to infrastructure without a shared con-
ceptual framework. This article develops a conceptual analysis to critically
interpret the existing literature, reviewing different infrastructure definitions,
classifications, and, measurement used across studies.
JEL: H10, H54, K40

1. Introduction

Moving essentially from Barro (1988) and Aschauer (1989),

many studies analysing the link between infrastructures and the eco-
nomic development have been realised.
On this field there is a broad spectrum of theoretical view-
points, some of them diametrically opposed to one another. A ge-
neral consensus is achieved around the idea that basic infrastructure
facilities are important features related to economic performance.
Apart from this main idea, opinions differ greatly: both magnitude
and causality remain subjects of debate.
Indeed, the seminal work of Aschauer (1989) estimated an
output elasticity of core infrastructure of .24 – i.e. a 1% increase in
investment in public infrastructure will result on a 0.24% increase in
the output of the private sector – so that this high elasticity led the
Author to argue that the decline in productivity growth during the
1970s was largely due to a decline in public investment in infra-
structure.
Nonetheless, as research in the field progressed, disputes
over this high impact of infrastructure arose. Gramlich (1994), for

100
example, pointed out that Aschauer (1989)’s approach was affected
by several drawbacks. In relation to the magnitude of infrastructure’s
impact, he highlighted that generally a positive public capital elas-
ticity forces the choice between increasing returns of scale and large
factors rent, and that Aschauer (1989)’s work result in “pretty stra-
tospheric estimates of the marginal product of government capital”
(Gramlich, 1994, p. 1186).
Moreover, the statistical causality between infrastructure
and productivity itself is questioned, with Looney and Frederiksen
(1981)’s words, one of the research question is as follows: “is
infrastructure the initiating factor in the development process or it is
merely a passive or accommodating factor?” (Looney and Frederik-
sen, 1981, p. 286)
At this regard, Evans and Karras (1994) – in their study
regarding seven OECD countries between 1963 and 1988 – even
founding strong correlations between the two variables, concluded
that the direction of causality was the opposite of that reported by
Aschauer (1989) – i.e. increased stocks of public capital were the
result of increased productivity and economic growth, not the cause
– arguing that “there is no evidence that government capital is highly
productive” (Evans and Karras, 1994, p. 278). As possible theo-
retical justification of this empirical result can be invoked Zegeye
(2000)’s argument that infrastructure is a normal good, so that
wealthy counties will tend to have more due to their higher level of
income.
Many other studies often sustain an intermediate thesis
distinguishing between (more or less) productive and unproductive
infrastructure and trying to deal with infrastructure endogeneity pro-
blem with appropriate econometric tests.
They could be grouped together into four approaches:
(i) The production function approach that models the
amount of output that can be produced for each factor of production,
given technological constraints. In this approach public infrastructure
enters as a free input furnished by government.
(ii) The cost function approach takes into account factor
prices such as the price of labour, machinery, and finance. Public
infrastructures are here conceived as costs saving factors.
(iii) Growth models belonging to the tradition of endoge-
nous growth are often augmented in order to consider as growth-
enhancing factors also public infrastructures.

 101
 (iv) Data-oriented models analyse relations between several
data series including infrastructures and GDP without relying heavily
on any particular economic theory.
However, approaching the theme regarding the link bet-
ween infrastructure and productivity, especially in empirical terms,
two important preliminary questions arise: what is infrastructure?
How to measure it?
Indeed, in the absence of standard definitions any com-
parison between studies is challenging: referring to “infrastructure”
various measures of road, electricity generating plants, water and
sewerage systems etc. have been utilised, often without a clear
statement of the criteria utilised to define what is infrastructure. In
addition, various types of measures (e.g. financial-flow, financial-
stock, physical) have been utilised in literature.
Many literature reviews have been realised concerning stu-
dies on infrastructures’ impact on productivity. For example, Infra-
structure Canada (2007), Sturm, Kuper et al. (1996), and Romp and
Haan (2007).
However, the issue of infrastructure’s definition, classifi-
cation, and measurement received less attention and most often is
treated in incidental way.
Bearing these issues in mind, this paper zooms in on infra-
structure definition (section 2), on its classification (section 3), and
on problems related to the measurement of infrastructure (section 4).
Section 5 presents some concluding remarks.

2. What is infrastructure?

There is no standard definition of infrastructure across eco-

nomic studies. Tinbergen (1962) introduces the distinction between
infrastructure (for example, roads and education) and superstructure
(manufacturing, agricultural and mining activities) without neither
precise definitions nor any theoretic references of these terms.
The reason for this unsatisfactory situation comes from the
need for simultaneous realization of three analytic objectives: (i) the
formulation of a concept for the term “infrastructure”; (ii) the incor-
poration of theoretic approaches (for example, the theory of public
goods), and (iii) the description of the reality of infrastructure pro-
vision.


102
 According to Buhr (2003), the broadest economic version of
the term “infrastructure” – referring to the works of List (1841) and
Malinowski (1944) – dates back to Jochimsen (1966)’s book on the
theory of infrastructure, in which the author aims to present pre-
paratory studies for a modern theory of the development of a market
economy based on the study of infrastructure endowment.
By dividing the relevant time-paths of economic develop-
ment in (a) quasi-stagnation, (b) economic dualism, and (c) self-
sustained development1 he denotes “infrastructures” as the impor-
tant preconditions of economic development concerning the time-
path mentioned above and the transformation processes leading from
one step to another; in this framework infrastructures are provided by
the state or controlled by it.
More deeply, the author defines infrastructure as

the sum of material, institutional and personal facilities

and data which are available to the economic agents and
which contribute to realizing the equalization of the re-
muneration of comparable inputs in the case of a suitable
allocation of resources, that is complete integration and
maximum level of economic activities (Jochimsen, 1966,
p. 100).

Or, in a pragmatic sense, material infrastructure is under-

stood as
[…] 1. the totality of all earning assets, equipment and
circulating capital in an economy that serve energy pro-
vision, transport service and telecommunications; we
must add 2. structures etc. for the conservation of natural
resources and transport routes in the broadest sense and
3. buildings and installations of public administration,
education, research, health care and social welfare (Joch-
imsen, 1966, p. 103).

However, Jochimsen (1966)’s definition, as noted by Buhr

(2003), “has the disadvantage of not making factor price equalization
concrete” (Buhr, 2003, p. 1). A second fault of this definition is that
it “understands material infrastructure to be an enumeration of essen-
tially public facilities characterized by specific attributes” (Buhr,
2003, p. 1). Indeed, in Buhr (2003) the mainstream approach based

 103
on infrastructure attributes is totally rejected in favor of an approach
based on infrastructure specific functions (see further on this section).
Therefore, due to the absence of a standard (precise) defi-
nition, various authors model a variety of different indicators of
infrastructure and this fact, in turn, makes challenging any com-
parison involving different studies.
In addition, in terms of policy, having no common defi-
nition of infrastructure makes it difficult developing uniform policies
in this field (Infrastructure Canada, 2007).
Despite these difficulties related to its exact meaning, in the
public discussion, the term made a successful terminological career,
rising to a formula of political technocracy so that we “have” to
confront the issue of its definition.
Aiming to highlight general features of “goods” utilized
from time to time, we can say that the term “infrastructure” –
stemming from the usage of military language (where it refers to
permanent military installations such as barracks and airports) – in
economic sense fills two main criteria: i) infrastructure is a capital
good (provided in large units) in the meaning that it is originated by
investment expenditure and is characterised by long duration, tech-
nical indivisibility and a high capital-output ratio; ii) infrastructure is
also a public (sometimes a merit) good, not necessarily in the sense
that it is owned by the public sector, rather in the proper economic
sense that it fulfils the criteria of being not excludable and not rival
in consumption, for which economic agents show real (in the case of
merit goods) or opportunistic (in the case of public goods) “wrong”
(down-biased) preferences. Sometimes the characteristic of being a
public good is “weakened” so that infrastructure does create external
effects but do not achieve the maximal level of externalities re-
presented by public goods.
As mentioned above, the approach based on technical,
economic and institutional infrastructure features (Youngson, 1967;
Biehl, 1986) could be considered the main stream approach.
Nevertheless, an alternative approach has been developed
based on infrastructure essential2 functions: the so-called “functional
approach”.
The starting point of this second one is represented by the
idea that the creation of the social product is due to economic agents
interacting with each other and that the contribution of each agent is
based on the provision of infrastructures. Put differently, the peculiar


104
characteristic of the term “infrastructure” should be individuated
both in the activation and in mobilisation of the economic agents’
potentialities.
Therefore, according to this approach, material infrastruc-
ture, for example,
has the function of rendering possible the opening and
development of the economic agents’ activities. It puts
into action the potentialities of economic units for the
benefit of society (Buhr, 2003, p. 13).

Hence, each type of infrastructure can be defined according

to its effects. So that, for example, market-oriented (material) infra-
structure could be defined as all capital goods serving the coordi-
nation and interaction of economic units to realise their economic
plans.
Following this alternative approach to the issue of infra-
structure definition – i.e. the functional one – Buhr (2003) defines
infrastructure as “the sum of all relevant economic data such as rules,
stocks, and measure with the function of mobilising the economic
potentialities of economic agents” (Buhr, 2003, p. 16).
To summarise: this section presented two different general
definition of infrastructure based respectively on its attribute and on
its functions.
Next section will focus on different infrastructure classifi-
cation introduced in literature with the purpose to better define the
borders of this “elusive” term.

3. Infrastructure classification

Once introduced, in previous section, a general definition of

infrastructure, this section considers the different ways in which
infrastructures have been classified by different authors. In what
follows I will briefly describe the criteria used in literature to iden-
tify the categories of personal, institutional, material and immaterial
infrastructures; economic and social infrastructures, as well as core
and not-core, basic and complementary, network, nucleus and ter-
ritory infrastructures.
As key to an understanding of this classification should be
noted that classifications developed here are potentially overlapping,

 105
in the sense that, for instance, “roads” belong to material-economic-
network infrastructures according to the different point of view of the
analysis (on that point see table 2).
Personal, institutional and (im)material infrastructures. To
begin with, I will take into account Jochimsen (1966)’s distinction
between material, personal and institutional infrastructures.
I prefer describing personal and institutional infrastructure
first, in order to develop more in detail the material one.
Personal infrastructure refers to “… the number and the
qualities of people in the market economy characterized by the
division of labour with reference to their capabilities to contribute to
the increase of the level and the degree of integration of economic
activities” (Jochimsen, 1966, p. 133).
A general way to refer to personal infrastructure is repre-
sented by human capital defined by the Organization for Economic Co-
operation and Development (OECD) as

the knowledge, skills, competencies and attributes em-

bodied in individuals that facilitate the creation of per-
sonal, social and economic well-being” (OECD, 2001, p.
18).

So that the concept of human capital entails

a) the capacity of interpreting flows of sensory data and
structured information required for purposive individual
actions and inter-personal transactions among economic
agents;
b) the capacity for providing a variety of physical labour
service-inputs in ordinary production processes;
c) the cognitive basis of entrepreneurial market activities;
d) the key resource utilised for managing market and
non-market production, as well as household consump-
tion activities;
e) the creative agency in the generation of new know-
ledge underlying technological and organisational inno-
vations.” (David, 2001, p. 19)

As Buhr (2003) pointed out, the role of personal infrastruc-

ture for determining the quality of the economic agents’ values
(achievement motivation, productive capacity, value integration) re-
sults in three main aspects: (a) the tasks of economic agents in the


106
economic process (entrepreneurial guidance, unskilled and qualified
labour, teaching etc.), (b) the importance of personal infra- structure
for the individual (short-term and long-term consumption of edu-
cation), and (c) the social relevance of personal infrastructure (inte-
gration effect of education).
Institutional infrastructure “comprises the grown and set
norms, institutions and procedures in their reality of constitution,
insofar as it refers to the degree of actual equal treatment of equal
economic data, excluding meta-economic influences. It determines
the framework within which economic agents may formulate their
own economic plans and carry them out in co-operation with others”
(Jochimsen, 1966, p. 117).
In the sense introduced above, institutional infrastructure
stems from the term “economic constitution” and can be considered
the real implementation of the norms in the “institutional basis” of
the market economy. Thus, institutional infrastructure, being as-
signed the function of social integration of values, is the object of
economic and legal policy.
Let us turn to the definition of material infrastructure.
Given an economic setup (population preferences, the level of tech-
nology, the institutional rules, level of development, and, geogra-
phical particularities of a community) material infrastructure is
essentially characterized by two distinguishing qualities: i) fulfil-
ment of social needs and (economic necessity of) ii) mass pro-
duction.
The first attribute refers to the essential needs of human life.
Following this perspective, material infrastructures can be defined as
goods and services able to satisfy those wants of economic agents
originating from physical and social requirements of human beings.
For example, the need of drinking water is met by the
corresponding supply of water collected, say, in a reservoir which, as
a capital good, is a specific type of material infrastructure.
The output relative to a material infrastructure results from
the interplay of its corresponding supply and demand depending on
physical or social wants.
The supply side depends on production functions, finance
situation, and organizational structures of infrastructure producers,
such as industrial enterprises and administrative units. As general
rule it can be said that the production functions relate infrastructure
outputs to the factors of production. Yet, in other cases (e.g. in the

 107
 case of roads) infrastructure outputs are related to the direct uti-
lisation of capital stocks over time as a result of preceding pro-
duction processes.
With respect to the demand side, the different requirements
of human life to be satisfied by material infrastructure could be
driven – without any pretension of completeness – from the first
column of table 1 taken from Buhr (2003) which reports also in-
frastructures devoted to their satisfaction.

Table 1. Material infrastructure to satisfy requirements of human life.

Source: Buhr (2003), p. 22.
Want Infrastructure output Material infrastructure
(good or service)
Physical requirements
Water drinking water, water for reservoirs, canals, waterways, pipes,
industrial uses, irrigation water, irrigation facilities
water for generating hydro-
electric power
Warmth gas, oil, electricity, coal, nuclear drilling platforms, pipelines,
energy generation plants, coal mines
Light electricity, gas generation plants, drilling plants,
circuits, pipelines
Health medical care, refuse collection, hospitals, dumps, sewerage systems
waste water disposal
Protection against accommodation, working places, houses, buildings, plants, levees
nature, shelter flood protection
Social requirements
Security legislation (laws), judiciary, public buildings, police stations,
stability of the value of money, military installations
protection against crimes,
outward defense, military goods
Information usage of telephones, mobile telecommunication facilities, post
phones, radios, television, offices, newspaper production works
Internet, newspapers
Education child care, lectures, research, kindergartens, schools, universities,
lending out books research institutions, libraries
Mobility usage of roads by cars, buses, roads, highways, tracks, train
trucks, usage of tracks by trains, stations, airports, ports
usage of airports by airplanes,
usage of ports by ships
Environmental clean air and water air purification filters, waterworks
protection


108
 From what stated above should be clear that material infra-
structure facilities are usually highly complementary to each other.
As an example for all could be considered housing in relation to
public utility networks (e.g., water and energy supply equipment).
The second peculiar feature of material infrastructure cited
above is the non-availability of infrastructure goods and services to
the individual household or firm for production and cost reasons, i.e.,
economic necessities of mass production. The usually high fixed
costs of facilities – generating economies of scale – require the
(often joint) production of large volumes of outputs.
Moreover, since the fixed costs are very different com-
paring various capital stocks, material infrastructure provision takes
place under the conditions of different market structures ranging
from the prevalent form of (natural) monopoly (e.g., electricity sup-
ply), to competition (e.g., housing construction).
In conclusion about material infrastructures, they can be
defined as
those immobile, non-circulating capital goods that essen-
tially contribute to the production of infrastructure goods
and services needed to satisfy basic physical and social
requirements of economic agents and unavailable to the
individual economic agents [...] for production and cost
reasons so that mass production is economically cogent
(Buhr, 2008).

In literature it is also frequent the use of immaterial infra-

structure (by contrast to material infrastructure) in order to indicate
some kind of infrastructure – primarily innovation and education
infrastructures – linked to the development of the material one as in-
tended above, for instance, research centres, innovation networks,
services to the enterprises, etc.
Economic and social infrastructures. Hansen (1965) distin-
guishes the infrastructures into economic and social according to the
fact that they act on the level of economic development of a territory
in direct or indirect way. The result of this point of view consists in
the division of local public overhead capital (OC) into
two components, “social” overhead capital (SOC) and
“economic” overhead capital (EOC). […] Those items
classified as EOC are primarily oriented toward the
support of directly productive activities or toward the

 109
 movement of economic goods. SOC items […] may also
increase productivity, the way in which they do so is
much less direct than in the case for EOC items (Hansen,
1965).

Thus, economic infrastructures directly support productive

activities; they are: roads, highways, airports, naval transport, sewer
networks, aqueducts, networks for water distribution, gas networks,
electricity networks, irrigation plant and structures dedicated to the
commodities transfer.
While social infrastructures, are those finalized to increase
the social comfort and to act on the economic productivity; they are:
schools, structures for public safety, council flat (not referable to
expenses of economic nature), plant of waste disposal, hospitals,
sport structures, green areas, and so on (Hansen, 1965).
Core and not-core infrastructures. It was said above that
Aschauer (1989) attributed a conclusive role to the public capital for
the economic growth of a country, particularly to the component of
the cores infrastructure.
The cores infrastructures include, for the Author, roads and
highways, airports, public transport, electric and gas networks, net-
work for water distribution and sewer networks. The not-core infra-
structures are a residual component (Aschauer, 1989).
The same type of classification is adopted in Mastromarco
and Woitek (2004) in which the public capital is expressly separated
into core and not core component, and where it is empirically under-
lined the role that every component assumes in determining the
different degree of development of the Italian regions of the Centre-
north in comparison to those belonging to the southern part of the
country.
Sturm, Jacobs et al. (1995) use also a similar distinction
between basic and complementary infrastructure. Where basic infra-
structure refers to main railways, roads, canals, harbours and docks,
the electromagnetic telegraph, drainage, dikes, and land reclamation
as opposed to complementary infrastructure category which includes
light railways, tramways, gas, electricity, water supply, and local
telephone networks.
Network, nucleus and territory infrastructures. In Biehl
(1991) a distinction emerges among network infrastructures and
nucleus infrastructures. The first ones referring to roads, railroads,
“water’s highway”, networks of communication, systems for energy

110
 and water provisioning; while the nucleus infrastructures, referring
to schools, hospitals and museums, are relatively characterized by an
elevated degree of immobility, indivisibility, “not-interchangeabi-
lity” and multi-purpose features.
This last distinction recalls another aspect typical of the
nucleus or punctual infrastructures, tied up to their ability of attrac-
tion. According to this last criterion network infrastructure are those
for which users’ geographical area of provenience coincides with the
territorial unity in which the infrastructure is located, or is possible
hypothesising that its ability of attraction is close to zero. Thus, it is
(rather should be) diffused in capillary way on the territory.
Finally, territory infrastructures include services that, even
if object of private investments and activities, have effects on the
territory attractiveness, on its quality of life and on the dynamics of
development.
Table 2 aims to summarise the different ideas about infra-
structure classification introduced above.
Table 2. Infrastructure classification.
Hansen (1965) Aschauer Sturm, Jacobs et al. Di Palma, Biehl (1991)
(1989) (1995) Mazziotta et
al. (1998)
Economic Core Basic Material Network
roads, highways, roads, (main) railways transport roads,
airports, naval highways, and roads, canals, network , railroads,
transport, sewer airports, public harbors and docks, water water-
networks, transport, electromagnetic system, highway,
aqueducts , electricity telegraph, drainage, energy networks of
networks for networks, gas dikes, land network communi-
water distribution, networks, reclamation. cations
gas networks sewers systems for
electricity network network energy and
irrigation plant, water
structures provisioning
dedicated to
commodities
transfer
Social Not-core Complementary Immaterial Nucleus
schools structures residual light railways, structures schools,
for public safety component tramways, gas dedicated to hospitals,
council flat, plant networks, develop- museums
of waste disposal, electricity network, ment,
hospitals, sport water supply, local innovation
structures, green telephone network and
areas. education

 111
 Turning on the empirical side it is worth taking into account
how official statistics address the theme of infrastructure clas-
sification. At this regard, the following scheme illustrates the
composition of the macro-areas divided into areas and sub-areas
according to Italian ISTAT’s classification.
Table 3. Infrastructure classification according to macro-area area and sub-
area. Source: ISTAT (2006, p. 16), author’s translation.
Economic infrastructures
Transport Network road transport, railway transport, air transport,
sea transport.

Energy Network electricity network gas network water-system.

Social Infrastructures
Health Infrastructures free hospital treatment, health service, social
security.

Educational nursery, primary, school for pupils aged 11–14,

Infrastructures secondary school, compulsory education,
universities.

Cultural Infrastructures cultural, artistic and historic heritage, theatre,

music, cinema and entertainment, Sport.

Environmental water purification plant, Waste disposal, Green

Infrastructures areas.
Territory Infrastructures
Tourist Infrastructures tourist receptiveness.
Trade Infrastructures retail trade, wholesale trade.
Monetary
Intermediation monetary intermediation.
Infrastructures
Note: for each entry in the right side of the table a residual sub-area labeled “other
aspects” which is reported in the original table is suppressed.

Therefore, economic infrastructures include areas related to

networks for commodities and people transport, those for the energy,
water, and, gas transportation.


112
 The macro-area related to social infrastructures comprises,
in turn, four areas: the infrastructures of the health, education,
culture and of the environment infrastructures.
The last macro-area concerns the territory infrastructures
and includes resources for commerce, tourism and for monetary in-
termediation.
Once defined infrastructure and presented some classifi-
cation introduced in literature, next section will focus on the problem
of its measurement.

4. How to measure infrastructure?

Preliminarily, should be noted that the goal of infrastructure

measurement is essentially twofold.
First, one could be interested in calculating a measure of
infrastructure that aims to quantify the existing infrastructure in
order to insert it into the national statistical system.
Second, one could be interested in obtaining a measure of
infrastructure with the purpose to investigate its effects in terms of
(competitiveness and) development of a territory (Brancalente, Di
Palma et al., 2006).
Certainly, each category of infrastructure introduced in sec-
tion 3 presents peculiar difficulties related to both purposes.
For example, the measurement of institutional infrastructure
goes deeply in the character of civic life – involving political
stability, quality of government, and, also social infrastructure – so
that its exact “measurement” is rather ambitious.
Another significant example is constituted by human capital
representing a crucial factor in endogenous growth models and
widely used despite difficulties regarding its measurement.
At this regard it is worth noting (without providing details)
that while Easterly and Rebelo (1993) included “two school en-
rolment variables […] as proxies for the initial level of human
capital” (Easterly and Rebelo, 1993, p. 424), Marrocu, Paci et al.
(2005), with the same purpose, use a 1996–2002 average of public
spending on various categories; namely: education, training, research
and development, pension, wage supplementation, and, labour.
In order to generalise across studies and categories of in-
frastructures could be said that in literature the problem of infra-
structure quantitative analysis has received two main different
 113
solutions: the first measuring the level of infrastructure endowment
in monetary terms, the second measuring it in physical terms.
Furthermore, a second sub-distinction involving both me-
thods can be operated.
In monetary terms, infrastructure may be intended as a flow
or a stock variable. In the first case, (government) spending cor-
respond to the provision of public services that instantaneously affect
the production. In the second case, instead, what government spends
“today” is added to the stock of public capital and affects the future
production process (Irmen and Kuehnel, 2008).
Typically, in order to calculate the stock measure of infra-
structure from financial flow, researchers use the perpetual inventory
method (PIM) which consists in adding up past gross3 investments,
adjusted for depreciation (see appendix A for details).
As noted above, both methods have been utilised in liter-
ature. For instance, Barro (1988) used productive government ex-
penditure as a flow variable and after his seminal work many studies
have done similarly, among others Everaert and Heylen (2004);
Ghali (1998; Everaert and Heylen (2004; Belloc and Vertova (2006);
Mittnik and Neumann (2001); Pereira (2000; Pereira (2001).
Infrastructure is instead considered as stock variable, for
example, in Albala-Bertrand, Mamatzakis et al. (2004); Bonaglia, La
Ferrara et al. (2001); Ferrara and Marcellino (2000), and Kamps
(2006).
When infrastructure is considered in physical terms – es-
pecially with respect to material infrastructures – two variations are
possible. The physical endowment can be considered simply in phy-
sical terms (e.g. kilometres of roads, electrical generating capacity,
number of hospitals, etc.) or can be measured the physical endow-
ment and then transformed in monetary terms attributing a price to
each category of good, that is adopting the so-called common in-
ventory method (CIM) .
It is worth stressing that, as Brancalente, Di Palma et al.
(2006) noted with respect to the Italian case, adopting one or the
other approach leads to results that could differ greatly.
Moreover, the difference between the two measures in-
creases with the territorial detail of the analysis. In particular, com-
paring two studies utilising the physical approach (Di Palma and
Mazziotta, 2002; Istituto Guglielmo Tagliacarne, 1998) and two
studies adopting the PIM (Montanaro, 2003; Picci, 1995) the authors


114
find that the strong regional north-south disparities reported in both
works regarding the physical approaches disappear in the works
using the PIM.
From the perspective of the analysis aiming to study the
infrastructure’s impact on productivity this variety of methodologies
– potentially leading to significantly different results one from the
other – raises the opportunity to consider critically each method in
order to assess its advantages and disadvantages that should con-
sidered when interpreting the results of the analysis as a whole.
First, as Romp and Haan (2007) noted, with regard to the
financial side one should be aware that in “applying the […] per-
petual inventory method, the researcher has to make certain assump-
tion about the assets’ lifespan and depreciation. Furthermore, one
needs an initial level of the capital stock. Especially with infra-
structure these assumption are far from trivial” (Romp and Haan,
2007, p. 13).
Furthermore, Brancalente, Di Palma et al. (2006) argue that
the same concept of withdrawing is debatable when applied to
(public) infrastructure. Indeed, the authors argue that, while it is
reasonable to think about withdraw concerning industrial machine
and various equipment owned by private sector especially for those
subject to rapid technical obsolescence, it is not the same in the
public infrastructure case: roads, bridges, ports.
Second, Pritchett (1996) argued that due to (in)efficiency or
structural reasons, public stock based on investment series will tend
to be over-evaluated.
While Montanaro (2003), focusing on the Italian case, attri-
butes the difference between the financial and the physical side to
morphology, population density and inefficiency, Golden and Picci
(2005), once tested the statistic (in)significance of the first two fac-
tors, attribute the difference to corruption since, they argue, cor-
ruption and inefficiency are strictly (rather perfectly, as implicitly
assumed in their paper) correlated.
Moreover, public investment series itself depend heavily on
the definition of public sector adopted by the national account sys-
tem. It is worthwhile referring once more to Romp and Haan (2007)
citing the piece in which they noted that thinking about infrastructure

[m]ost people probably think about roads and other in-

frastructure – such as electricity generating plants and
water and sewage systems – when they refer to the
 115
 public capital stock. However, it is important to point
out here that this does not fully correspond to the con-
cept of public sector Investment expenditure as defined
in national accounts statistics, which are typically used
to construct data on public capital stock. [Because] only
spending by various government sectors is included.
That implies that spending by the private sector (in-
cluding public utility firms concerned with electricity
generation, gas distribution, and water supply) is ex-
cluded (Romp an Haan, 2007, p. 13).

Third, simply adding up past investment do not take into

account that the effects of public investment might depend on the
level of corresponding capital stock (Kamps, 2006).
Fourth, from a network perspective PIM values have certain
pitfalls: the internal composition of the stock matters, since the
marginal productivity of one link depend on the capacity and con-
figuration of all links in the network. Using measures of total stock
may thus allows one to estimate the average marginal product of
road, say, in the past, but these estimates may not be appropriate for
considering the marginal product of additional roads today (Fernald ,
1999).
Finally, on the strict computational side, PIM requires long-
term time series on public investment and this type of data are not
always available for all country and for all level of government. It
does exist data for most OECD countries, but for many developing
countries public stock infrastructure cannot be constructed.
With regard to the physical side, in general terms, can be
said that this kind of measure have been employed in order to deal
with the most part of problems arising from PIM, see Canning and
Pedroni (1999), Sanchez-Robles (1998), and, Esfahani and Ramìres
(2003).
In fact, the measures utilised – such as number of kilo-
metres of paved roads, kilowatts of electricity generating capacity,
number of telephones line and so on – have the advantage that they
do not rely on the concept of public investment as employed in the
national accounts and, in addition, some of the measures do not
necessarily refer to (the results of) government spending.
However, we still need a measure strictly related to some
instrument in terms of policy (first of all public spending) and, per-


116
haps more important, simple physical measures do not correct for
quality which is a crucial point in infrastructure effectiveness.
For instance, in ISTAT (2006) the indicators, let say in the
area of the sanitary infrastructure, report the availability of hospitals
and beds for each of the various specializations or, in the area of the
educational infrastructures, the availability of scholastic buildings
and classrooms without any information about the quality of such
elements. Moreover, the quality in itself is also difficult to measure.
Coming back to the main distinction between monetary and
physical measurement treated in this section, should be noted that
although the financial and the physical approach potentially produce
completely different measures, the two approaches could be com-
bined in view to draw important conclusions, for example, about the
return rate of public expenditure, in different areas of the country.

5. Concluding remark

Many studies utilise the term “infrastructure” with par-

ticular respect to its social and economic impact. Nonetheless, it does
not exist a standard definition of the term. Hence, many goods have
been labelled as infrastructure according to various classifications
and with different techniques of measurement. This circumstance
makes it challenging any comparison.
This paper aimed to provide a general framework of ana-
lysis regarding infrastructure’s definition and the related issue of its
classification and measurement.
In so doing, this paper makes an original contribution to
already existent literature reviews on infrastructure to the extent that
it represents an attempt to critically illustrate difficulties arising in
answering to the question “what is infrastructure?” according to the
massive literature on infrastructures.
Therefore, this work started posing the main question of
infrastructure definition concluding that a precise definition is dif-
ficult to achieve because of difficulties in simultaneously achieving
three main objectives. Namely, formulating a “concept” for the term
infrastructure, incorporation of theoretic approaches, and, the des-
cription of empirical evidence of infrastructure provision.
Despite these difficulties, two main characteristics of infra-
structures – i.e. being a (i) capital (ii) public good – are presented at
the end of section 2.
 117
Once dealt with the issue of infrastructure definition a review of
different infrastructure classification is presented showing that, ge-
nerally, various categories of infrastructures are overlapping.
This fact could be read as an additional source of ambiguity,
since referring to the same good, scholars could refer their analysis
to different infrastructures categories.
Translated in terms of policy this evidence is not irrelevant.
Indeed, since often measures, rather than being good-based, are sec-
tor-based attributing result to one or the other category will result in
different policies.
The problem of infrastructure measurement considered in
the final part of this article is often an underlying issue of studies
developed in this field. However, each of the four different ap-
proaches to infrastructure measurement (financial-flow, financial-
stock, physical, and transformation of physical measure into mone-
tary values) is potentially leading to different values, and in turn, to
different empirical results.
While both the problem of infrastructure definition and
classification could not have a unique solution (the “best” solution
depending on authors’ preference and purposes), regarding the pro-
blem of measurement Brancalente, Di Palma et al. (2006) argued that
for (national) accountability purposes the preferable approach is the
monetary one, given that the whole framework is characterised by
monetary values. By contrast, aiming to study the impact of infra-
structure (for example by means of a production function approach)
a physical-based measure should rend one more confident about
results achieved to the extent that such a measure is able to better
represent the real infrastructure endowment of the economic system
under analysis.
Nevertheless, it should be noted that even if the two ap-
proaches – monetary and physical – are, in general, “neither con-
vergent nor compatible” (Brancalente, Di Palma et al., 2006, p. 265),
the possibility to use them in a complementary way, as in Montanaro
(2003) and Golden and Picci (2005) with respect to the Italian case,
is not precluded.
On the contrary, it represents a promising approach able to
increase analyses’ accuracy.


118
 NOTES

1. Where quasi-stagnation is characterized by a relatively constant

level of economic activities, mostly the subsistence level, due to the absence of
any stimuli to change; dualism results in the disintegrating decomposition of the
economy into segments with differently changing activity levels with respect to
sectors, regions and firm sizes due to the linkages of external effects, institutional
rigidities, technological discontinuities and other frictions of the market economy
and self-intensifying growth, is characterised by an increasing level of economic
activities.
2. The term “essential” refers to the fact that infrastructure initiate the
changes of economic variables.
3. For the rationale for using gross investment see Alvaro, G.,
(Ed.), (1999). Contabilità nazionale e statistica economica. Bari, Cacucci.

REFERENCES

Albala-Bertrand, J. M. et al. (2004), “The Impact of Public Infra-

structure on the Productivity of the Chilean Economy”, Review of Deve-
lopment Economics 8: 266–278.
Alvaro, G. (Ed.) (1999), Contabilità nazionale e statistica eco-
nomica. Bari: Cacucci.
Aschauer, D.A. (1989), “Is Public Expenditure Productive?”,
Journal of Monetary Economics 23: 177–200.
Azariadis, C. and Drazen, A. (1990), “Threshold Externalities in
Economic Development”, Quarterly Journal of Economics 105: 501–526.
Barro, R. J. (1988), “Government Spending in a Simple Model of
Endogenous Growth.” National Bureau of Economic Research Working Pa-
per No. 2588.
Becker, G., Murphy, K. et al. (1990), “Human Capital, Fertility,
and Economic Growth”, Journal of Political Economy 98: 12–37.
Belloc, M. and Vertova, P. (2006), “Public Investment and Eco-
nomic Performance in Highly Indebted Poor Countries: An Empirical Ass-
essment,” International Review of Applied Economics 20: 151–170.
Biehl, D. (1991), “The Role of Infrastructure in Regional Deve-
lopment,” in Vickerman, R. W., (Ed.), Infrastructure and Regional Deve-
lopment. London: Pion.
Bonaglia, F., La Ferrara, E. et al. (2001), “Public Capital and
Economic Performance: Evidence from Italy,” IGIER Working Paper No.
163.
Brancalente, B., Di Palma, M. et al. (Ed.) (2006), Investimenti,
Capitale Pubblico e Dotazione Fisica di Infrastrutture nelle Regioni Italiane
Federalismo, equità, sviluppo. I risultati delle politiche pubbliche analizzati
e misurati dai conti pubblici territoriali. Bologna, Il Mulino.

 119
 Buhr, W. (2003), “What Is Infrastructure?”, Department of Eco-
nomics, School of Economic Disciplines, University of Siegen. Siegen Dis-
cussion Paper No. 107.
Canning, D. and Pedroni, P. (1999), “Infrastructure and Long Run
Economic Growth”, mimeo.
Di Palma, M., Mazziotta, C. et al. (Eds.) (1998), Infrastrutture e
sviluppo. Primi risultati: indicatori quantitativi a confronto (1987–95).
Quaderni sul Mezzogiorno e le politiche territoriali. Roma: Confindustria.
Easterly, W. and Rebelo, S. (1993), “Fiscal Policy and Economic
Growth. An Empirical Investigation”, Journal of Monetary Economics 32:
417–458.
Esfahani, H. and Ramìres, M. T. (2003), “Institutions, Infrastruc-
ture and Economic Growth”, Journal of Development Economics 70: 443–
477.
Evans, P. and Karras, G. (1994), “Is Government Capital Pro-
ductive? Evidence from a Panel of Seven Countries”, Journal of Macro-
economics 16(2): 271–279.
Everaert, G. and Heylen, F. (2004), “Public Capital and Long-
Term Labour Market Performance in Belgium”, Journal of Policy Modelling
26: 95–112.
Fernald, J. (1999), “Assessing the Link between Public Capital
and Productivity”, American Economic Review 89: 619–638.
Ferrara, E. L. and Marcellino, M. (2000), “TFP, Costs, and Public
Infrastructure: An Equivocal Relationship,” IGIER Working Paper No. 176.
Ghali, K. H. (1998), “Public Investment and Private Capital For-
mation in a Vector Error-Correction Model of Growth”, Applied Economics
30: 837–844.
Golden, M. A. and Picci, L. (2005), “Proposal for a New Measure
of Corruption Illustrated with Italian Data”, Economics and Politics 17(1):
38.
Goldsmith, R. W. (1953), “A Perpetual Inventory of National
Wealth”, National Bureau of Economic Research 14.
Gramlich, E. M. (1994), “Infrastructure Investment: A Review
Essay”, Journal of Economic Literature 32(3): 1176–1196.
Hansen, N. M. (1965), “The Structure and Determinants of Local
Public Investment Expenditures”, Review of Economics and Statistics 2:
150–162.
Infrastructure Canada (2007), “Infrastructure and Productivity: A
Literature Review”, Transport Infrastructure and Communities Portfolio,
Government of Canada. Research and Analysis Division Infrastructure.
ISTAT (2006), Le Infrastrutture in Italia. Un'analisi della dota-
zione e della funzionalità. Rome.
Jochimsen, R. (Ed.) (1966), Theorie der Infrastruktur: Grundlagen
der marktwirtschaftlichen Entwicklung. Tübingen: J.C.B. Mohr.


120
 Kamps, C. (2006), “New Estimates of Government Net Capital
Stocks for 22 OECD Countries 1960–2001”, IMF Staff Papers 53: 120–150.
Krueger, A. (1997), “Experimental Estimates of Education Pro-
duction Functions”, NBER Working Paper No. 6051, Massachusetts, June.
List, F. (Ed.) (1841), Das Nationale System der Politischen Öko-
nomie. Stuttgart: Cotta Verlag.
Looney, R. and Frederiksen, P. (1981), “The Regional Impact of
Infrastructure Investment in Mexico”, Regional Studies 15(4): 285–296.
Lucas, R. (1988), “On the Mechanics of Economic Development”,
Journal of Monetary Economics 22: 3–42.
Malinowski, B. (Ed.) (1944), A Scientific Theory of Culture and
Other Essays. Chapel Hill, N.C.: University of North Carolina Press.
Marrocu, E., Paci, R. et al. (2005), “Glie effetti del capitale pub-
blico sulla produttività delle regioni italiane”, CRENoS Working Paper No.
200201, University of Cagliari and Sassari, Sardinia.
Mastromarco, C. and Woitek, U. (2004), “Public Infrastructure
Investment and Efficiency in Italian Regions”, www.cide.info/conf/papers/
1161.pdf.
Mittnik, S. and Neumann, T. (2001), “Dynamic Effects of Public
Investment: Vector Autoregressive Evidence from Six Industrialized Coun-
tries”, Empirical Economics 26: 429–446.
Montanaro, P. (2003), “Lo stock di capitale pubblico: una stima
per regione e per tipologia di bene”, Rivista economica del Mezzogiorno 3:
423–462.
OECD (1993), “Methods Used by OECD Countries to Measure
Stocks of Fixed Capital”, OECD.
Pereira, A.M. (2000), “Is All Public Capital Created Equal?”,
Review of Economics and Statistics 82: 513–518.
Pereira, A.M. (2001), “On the Effects of Public Investment on
Private Investment: What Crowds In What?”, Public Finance Review 29: 3–
25.
Pritchett, L. (Ed.) (1996), “Mind Your P’s and Q’s: The Cost of
Public Investment Is Not the Value of Public Capital”. Policy Reseach
Working Paper No. 1660. Washington, DC: World Bank.
Romp, W. and Haan, J. (2007), “Public Capital and Economic
Growth: A Critical Survey”, Perspektiven der Wirtschaftspolitik 8 (Special
Issue): 6–52.
Sanchez-Robles, B. (1998), “Infrastructure Investment and Growth:
Some Empirical Evidence”, Contemporary Economic Policy 16: 98–108.
Sturm, J.-E., Jacobs, J. et al. (1995), “Productivity Impacts of
Infrastructure Investment in the Netherlands 1853–1913”, University of
Groningen. Research Institute SOM (Systems, Organisations and Manage-
ment) Research Report 95D3.

 121
 Sturm, J.-E., G. H. Kuper et al. (1996), “Modelling Government
Investment and Economic Growth on a Macro Level: A Review”, CCSO
Series No. 29.
Tinbergen (1962). Shaping the World Economy, Suggestions for
an International Economic Policy. New York: The Twentieth Century Fund.
Youngson, A. J. (Ed.) (1967), Overhead Capital. A Study in Deve-
lopment Economics. Edinburgh: Edinburgh University Press.
Zegeye, A. A. (2000), “U.S. Public Infrastructure and its Contri-
bution to Private Sector Productivity”, U.S. Department of Labor Bureau of
Labor Statistics (June).

APPENDIX A

The Permanent Inventory Method:

theoretical and methodological aspects

The Permanent Inventory Method (PIM) is the technique

used in order to achieve a stock measure of infrastructure installed
based on the flow of financial investment. This is the most common
way that statistical agencies measure public capital stock. Essentially,
it involves adding up past capital formation in constant prices while
deducting the value of assets as they reach the end of their service
life.
The idea underlying PIM is that the consistence of capital
stock for the good at a given year ( K t ) depends on what was spent
during the previous L years with a cumulative process in which the
expenditure of each year is added to the previous one. Where L is the
g good’s average live. A complete review on this method is available
on Goldsmith (1953)
PIM requires an evaluation of the consistence of the stock
in one basic year, that can be achieved cumulating the series of the
fixed gross investments along the period corresponding to the good’s
average life.
If we hypothesize a simultaneous exit (i.e. a capital created
in a certain year is withdrawn in bulk at the end of its economic life),
then the gross capital stock at a given year (t) can be expressed
according to the following equation


122
 L −1
K t = ¦ I t −i (A.1)
i=0

where It is the fixed gross investment in the t year. An alternative

method is represented by a gradual exit in which capital created in a
certain year is withdrawn gradually during the time of its economic
life.
Once obtained the benchmark for a certain year, the stock
during the following years is simply given by the equation

K t +1 = K t + I t +1 − I t − ( L −1) (A.2)

Thus, in words, first I find the consistence to the beginning

of the period using the equation (1), subsequently using the equation
(2) I add the new investments and I subtract the value of the good(s)
that have exhausted to the time t their life of L years (that is why it is
used the sub-index t - (L-1)). For a more elaborate formulation see
OECD (1993)

A graphic representation may be useful to explain the

mechanism.

Figure A.1 – Permanent Inventory Method

Obviously this is a gross measure based on the idea that

each good maintain its value substantially unaltered during its eco-
nomic life. A “sophisticated” measure of net stock, K tn – i.e. a
measure that takes into account the deterioration and the obsoles-
cence of each good – needs some hypothesis regarding the de-
preciation function. Such a measure might be obtained considering
that each good gradually lose its value during its economic life.

 123
 At this regard the National Accounting System adopts a
constant depreciation function that means that a constant fraction of
the instrumental good is consumed during every year of its economic
1 ). In formula this idea can be
life (i.e. the depreciation rate is δ=
L
expressed as follows

K tn = K tn−1 + I t − Dt (A.3)

Where Dt represents the depreciation in the t year. Moreover,

if we hypothesize, as the National account do, a linear depreciation
function we can express Dt as follows

1 L −1
Dt = ¦ I t −i
L i =1
(A.4)
Hence, we can write the (3) in the following form

1
K tn = K tn−1 + I t − K t −1 (A.5)
L


124