Modern Economy, 2011, 2, 438-445
doi:10.4236/me.2011.23049 Published Online July 2011 (http://www.SciRP.org/journal/me)
Economic Diversity and Employment Levels in Italian
Provinces over the Period 1951-2001: Does a Recurrent
Industrial Pattern Exist?
Alessandro Marra1, Vittorio Carlei2, A le ss an dro Crociata 2
1Università degli Studi “G. D’Annunzio” di Chieti e Pescara, Facoltà di Scienze Manageriali and Dipartimento di
Metodi Quantitativi e Teoria Economica (DMQTE); GRIF (Gruppo di Ricerche Industriali e Finanziarie) - Fabio
Gobbo, LUISS Guido Carli, Rome
2Università degli Studi “G. D’Annunzio” di Chieti e Pescara, Facoltà di Scienze Manageriali and Dipartimento di
Metodi Quantitativi e Teoria Economica (DMQTE); SEMEION, Rome.
Received April 4, 2011; revised May 50, 2011; acce pted June 3, 2011
Abstract
Economic diversity is crucial to explain how geographic areas grow and evolve: economic diversity facili-
tates the transmission of ideas (through knowledge spillovers) and the outsourcing of activities
(city-nurseries). In this paper we focus on these theoretical perspectives and examine the industry structure
and growth of Italian provinces. We use data on employment levels in 47 economic sectors in 103 provinces
and measure their degree of diversity over the period 1951-2001. By taking into account both provincial size
(i.e., employment levels) and temporal dimension we test whether economic initiatives benefit from being
localised in areas that are highly diversified from an industrial point of view. After confirming the relevance
of economic diversity we investigate the industry pattern that boosted growth in the largest (in terms of
number of employees) provinces in our sample. Our results confirm the Jacobs’ intuitions about how cities
evolve and contribute to debate by reducing the gap between some Jacobs’ hypotheses and major theses.
Keywords: Economic Diversity, Industrial Pattern, Provincial Level
1. Introduction
The aim in this paper is to examine the economic
structure and evolution of Italian provinces and test un-
der what circumstances and to what extent economic
diversity is positively associated with economic growth.
At the theoretical level we rely on two distinct but
strongly associated perspectives based on knowledge
spillovers and the interpretation of the city as a nursery.
According to several authors, and much empirical
evidence, the most important transfers of knowledge are
inter-industry transfers [1-3]. From this perspective the
variety and diversity of geographically proximate Indus-
tries, rather than industrial specialisation, promote inno-
vation and growth. At the same time, some authors pre-
dict that firms located in areas that are highly industrially
diversified should grow more quickly [4-8], for exam-
ple, consider a representative set of cities and Industries
over 18 years and show that the extent of diversity in the
manufacturing industries in 1970 was crucial for attract-
ing new or high-tech industries and creating the condi-
tions for these n ew industries to flourish.
It is evident that the arguments proposed by [1,5] are
essential to both these theoretical perspectives. Accord-
ing to this author cities are settlements whose growth is
generated consistently from their own local economies:
their pools of skills, manufactures and materials, at once
diverse and concentrated, provide the best conditions for
the birth and growth of entrepreneurial small firms and
an ever-increasing division of labour on the basis of
which “new work can be added to old” [9,1,10,5].
We investigate employment levels (for which data are
available on such a large time span) of Italian provinces
to measure their degree of economic diversity over the
period 1951-2001. Using an index of entropy that ac-
counts for both provincial size (in terms of number of
employees) and time dimension, we test whether eco-
nomic initiatives benefit from being localised in highly

A. MARRA ET AL.439
diversified (from an industrial point of view) areas. Note
in this respect that our analysis is speculative in the sense
that we rely on Jacobs’ intuitions about how cities evolve
and, de facto, make use of data collected at provincial
level.
Analytically, economic diversity can be linked to
growth based on three different interpretations of diver-
sity: as a source of knowledge spillovers, as the con cept-
tual basis for the city-nursery perspective (described
above), and as a portfolio strategy to spread economic
risk and, hence, protect a geographic area from external
shocks [11,12]. We intend to find out which of these
interpretations is more suitable to explain the association
between diversity and growth in Italian provinces. Ac-
cordingly we investigate the industrial pattern that
boosted growth and prosperity in the 20 largest provinces:
depending on the nature of the industries dominating the
local economic functions, and their degree of comple-
mentarity, we provide some qualitative insights support-
ing the city-nursery interpretation.
2. The theoretical Roots of the Debate on
Economic Diversity and Growth
Economic diversity refers to the variety of the industrial
activities within a certain geographic area. It contributes
to the growth and evo lution of cities: economic diversity
enabled by proximity facili tates the transmission of ideas
(through knowledge spillovers) and the outsourcing of
activities (the city-nursery). Several empirical studies
debate the fact that economic diversity is strongly asso-
ciated with increased output and greater productivity and
growth. [13], in a study of 79 American cities, find that
the level of economic diversity in 1880 substantially af-
fected output levels in 1890. [14] control for several
variables (growth and variability of industries, relative
composition of fast and slow growth industries, and rela-
tive mix of changing and stable industries) and show that,
over the period 1969-1985, the industry mix of econo-
mies and their level of growth are positively associated.
[15], on the other h and, suggests th at ge ograph ic ind u stry
concentration (rather than industry diversification) en-
ables knowledge spillovers among firms and local
growth, based on evidence from computer chip manu-
facturers in Silicon Valley. [16] show that the concentra-
tion of economic activities (as measured by the density
of employment) is positively correlated to GDP levels.
There is some empirical evidence of specialization and
intra-industry knowledge spillovers providing the basis
for metropolitan growth [17], with some studies [18]
considering the effects of both specialization and diver-
sity [19].
How to support theoretically such a mixed evidence?
Many authors consider that inter industry knowledge
exchange is more important than intra industry transfers
[20,1-3,21,22]. In diversified economies there will be
more interchange of different ideas, and this diversity is
conducive to urban growth. [1] compares the economic
structure and evolution of Manchester, a former textiles
specialised city, which eventually declined, and Bir-
mingham, a highly diversified city, which eventually
flourished. [2] suppo rts Jacobs by arguing that the diver-
sity of urban activities automatically promotes attempts
to apply or adopt in a particular field, technological solu-
tions already utilised in another sector.
Jacobs’ first intuition has been interpreted and elabo-
rated by scholars to mean that because cities bring to-
gether people from different fields and areas, this fosters
the transmission and, in turn, the cross-fertilisation of
ideas across different activities, which is crucial for ur-
ban development.
An example here is brassiere manufacture, which
evolved from the innovations made by dressmakers
rather from than the lingerie industry [21]. Also, it is
reported that the financial services industry grew out of
intersectoral knowledge spillovers; more specifically, in
New York, it developed from the recognition of grain
and cotton merchants that there was a need for national
and international financial transactions. Similarly, the
practice of equipment leasing was initiated by a food
processor in San Francisco (not by bankers). [20] dis-
cusses the spread of machine tools and describes how
inventions or innovations spill over into different Indus-
tries, while [2] provides a detailed account of the sectoral
source of inventions, estimating that some 70% of in-
venttions within a given sector are exploited outside that
industry.
In line with the city-nursery perspective, some authors
argue that firms located in areas that are highly industri-
ally diversified should grow more easily [22,23,4,24,1,
5-7]. [9] stresses this aspect affirming that: “Although it
is hard to believe, while looking at dull gray areas, or at
housing projects or at civic centers, the fact is that big
cities are natural generators of diversity and prolific
incubators of new enterprises and ideas of all kinds.
Moreover, big cities are the natural economic homes of
immense numbers and ranges of small enterprises”
[9:145].
That is, big cities are generators of diversity and incu-
bators of new ideas, and are the natural economic home
of several and various small firms. Jacobs emphasises the
relevance of the urban dimension for small entrepreneur-
rial activities, arguing that small firms operating in a
limited market, in many cases, require the presence of
several services and inputs in order to outsource certain
activities. In Jacobs’ words: “Typically (small enter-

A. MARRA ET AL. 440
n
j
prises) must draw on many and varied supp lies and skills
outside themselves, they must serve a narrow market at
the point where a market exist, and they must be sensitive
to quick changes in this market. Without cities, they
would simply not exist” [9:145].1
[21] compares New York and Pittsburgh in terms of
their size, industrial structure and rate of growth, arguing
that New York’s evolution is linked to a more abundant
supply schedule of entrepreneurship. In recognising that
large areas are more diversified than small areas, Chinitz
claims that diversified areas exhibit more stable growth
because their fortunes are not tied to the fortunes of a
few industries [21:281]. Likewise, [25] depict New York
City as an incubator of new business activities and em-
phasise the constant rise and flow of new businesses as
the main source of its de- -velopment.2
In a study of small towns and rural areas [4] recogn ise
their ability to provide various and more competitive
“facilities” for industrial activities. As already mentioned,
[8] investigate a representative group of cities and in-
dustries over the period 1970-1987, showing that diver-
sity in manufacturing at the beginning of the period was
a crucial factor in attracting new and h igh tech industries
to the area and boosting their development.
3. Measuring Economic Diversity and
Employment Levels
Thus, the idea that economic diversity facilitates the
transmission of ideas and the outsourcing of activities is
rooted in the theoretical and empirical economic litera-
ture. In this paper we provide an investigation to test
whether, under what circumstances and to what extent,
the diversity of economic activity explains how Italian
provinces grow and evolve, based on data on employ-
ment levels in 47 economic sectors between 1951 and
2001.
We exploit Shannon’s diversity index (Ht’i) to meas-
ure the degree of diversity. This index, also known as the
entropy index, is a statistics that measures the deviation
of a given distribution from complete concentration
(minimum entropy) or dispersion (maximum entropy).
The entropy index measures the diversity of the eco-
nomic structure against a uniform distribution of em-
ployment where the benchmark is an equi-proportional
distribution of employment among all industries. The
entropy index is used in numerous areas and typically is
calculated using the fo llowing formula:
1log
tj
j
H
ic pp

  



(1)
where pj is the relative proportion of employment in the
j-th economic sector, n is the number of sectors, c is an
arbitrary constant which determines the scale of meas-
urement and t is the relevant time interval. Under condi-
tions of absolute diversity, the maximum value of Ht’i
for the i-th province is obtained; minimum entropy or
complete specialisation is indicated if only one of the pi
sector equals 1 and the remainder are zero. We calculate
seven measures of entropy (H51’i, H61’i, H71’i, H81’i,
H91’i, H96’i and H01’i), one for each year in our dataset
(1951, 1961, 1971, 1981, 1991, 1996 and 2001)3.
We measure employment levels as number of em-
ployees in the i-th province at time t between 1951 and
2001: emp51i, emp61i, emp71i, emp81i, emp91i, emp96i and
emp01i. Use of employment levels is imposed by data
availability and suggested as being appropriate by [26]
according to whom employment growth is seen as par-
ticularly appropriate to explain the impact produced by
diversity. Knowledge spillovers, in effect, will facilitate
radical innovation and product innovation which, in turn,
will lead to the creation of new markets and more em-
ployment; similarly, according to the city-nursery per-
spective, cities will attract new entrepreneurial in itiatives
which strictly are accompanied by additional employees.
After normalising the values of t
H
i

DH i
between 0 and 1,
we compute a dynamic measure of entropy (k

) to
account for a specific hypothesis: diverse industrial pat-
terns generate effects over the time. The presence of di-
verse economic activities at a given initial time period (t
= 1) will directly generate more diverse economic ac-
tive- ties at t = 2 and will continue positively and indi-
rectly to affect the degree of economic variety at the fol-
lowing time periods (t = 3, t = 4, t = 5, etc.)]. Accord-
ingly DHk’i, for


1,kt, is given by the cumulative
sum for


7
DH i
1,t
7
t
of the Ht’i
1
k
t
H i




 (3)
We get seven DHk’i, one for each observed year. Re-
lying on the fact that the positive effects that economi-
cally diverse provinces produce are generally amplified
by the dimension of the local economy, we propose to
3We use the inverse of the Herfindahl index (H is, per se, a measure o
f
industry concentration) to find support for the ranking provided by the
entropy measure. As is well known, the Hti index is

2
1
100
n
tj
j
Hi p



1In his The Location of Economic Activity [22] gives the example o
f
p
rinting businesses, where typesetting used to be outsourced to com-
p
ositors, and garment manufacture which relied on certain specialist
services for particular steps in the process such as buttonhole making.
2[23] in Logic of British and American Industry describes how firms in
large cities find it cheaper to ‘farm out’ certain functions to take ad-
vantage of scale economies, while firms in small places do not have
this alternative and must do these
j
obs themselves.

(2)
where pj is the number of employees in the j-th sector, and nis the total
number of sectors. Most of the records in the rankings overlap (com-
p
are 1st and 2nd columns in table 5 in the Annex
)
.

A. MARRA ET AL.441

4
DDH i

DDH i
DH i 
k
DDH iempiDH i


DDH i
include the size of provinces k. Formally,
k are the weighted by the number of em-
ployees (empti): k
kt (4)
We estimate the association per province and over
time between the seven k

and the seven values of
employment (empti) at each time t. It is interesting to
notice that most Italian provinces show a positive and
very high correlation coefficient: it is higher than 0.90 in
64 provinces and over 0.50 in 74 provinces (see Figure
1): increased diversity is accompanied by increased em-
ployment.
Only 19 provinces have a negative and significant
(below –0.50) correlation coefficient (i.e., employment
level per province is inversely associated to the evolution
of its diversity).
The provinces that experienced a highly correlated
evolution of both degree of diversity and employment
levels are Milano (1st), Bergamo, Frosinone, Mantova,
Parma, Ascoli Piceno, Bari, Como, Teramo and Brescia
(see Table 1). Note that here we are not looking at the
mere correlation between economic diversity and pro-
vincial size at a given time and that the above group in
cludes not only large but also medium and small prov-
inces. While changes in diversity and employment levels
are negatively associated over the time for Isernia
(ranked last), Pavia, Genova, Trieste, La Spezia, Savona,
Gorizia, Vercelli, Alessandria and Verbano.
4. Economic diversity and the recurrent
pattern of economic functions
We also want to investigate empirically [9] claim that:
“To understand cities, we have to deal outright with
combinations or mixtures of uses, not separate uses, as
the essential phenomena” [9: 144]. That is, in order to
study geographic areas, we need to focus on the several
relationships among economic activities, their combina-
tion and the cro ss fertilisation processes th at o ccur with in
the local economic structure. We focus on the largest, in
terms of employment levels, provinces at 2001.
Figure 1. Frequency distribution of correlation between
DDH′ki and empti, Source: Own elaboration on ISTAT da-
tabase (2001).
Table 1. Correlation between DDH’ki and empti per prov-
ince, 1951-2001.
Agrigento -0.735Genova -0.983 Potenza 0.980
Alessandria -0.936Gorizia -0.958 Prato 0.971
Ancona 0.990Grosseto -0.871 Ragusa 0.932
Aosta -0.885Imperia -0.058 Ravenna 0.925
Arezzo 0.982Isernia NaN Reggio di Calabria-0.861
Ascoli Piceno0.997La Spezia -0.971 Reggio nell’Emilia 0.992
Asti 0.405L Aquila 0.940 Rieti -0.593
Avellino 0.891Latina 0.986 Rimini 0.969
Bari 0.997Lecce 0.996 Roma 0.985
Belluno 0.976Lecco 0.974 Rovigo 0.986
Benevento 0.467Livorno -0.781 Salerno 0.985
Bergamo 0.998Lodi 0.580 Sassari 0.996
Biella 0.599Lucca 0.994 Savona -0.965
Bologna 0.984Macerata 0.994 Siena 0.928
Bolzano-Bozen0.982 Mantova 0.998 Siracusa 0.975
Brescia 0.996Massa-Carrara -0.816 Sondrio 0.956
Brindisi 0.983Matera 0.982 Taranto 0.991
Cagliari 0.386Messina -0.368 Teramo 0.997
Caltanissetta-0.810Milano NaN Terni -0.837
Campobasso 0.966Modena 0.990 Torino 0.346
Caserta 0.995Napoli 0.976 Trapani -0.409
Catania 0.921Novara 0.948 Trento 0.903
Catanzaro 0.876Nuoro 0.987 Treviso 0.991
Chieti 0.996Oristano 0.941 Trieste -0.974
Como 0.997Padova 0.994 Udine 0.990
Cosenza 0.249 Palermo 0.755 Varese 0.976
Cremona 0.346Parma 0.998 Venezia 0.989
Crotone 0.776Pavia -0.990 Verbano-C-O-0.924
Cuneo 0.950Perugia 0.995 Vercelli -0.951
Enna 0.438Pesaro e Urbino 0.988 Verona 0.991
Ferrara 0.645Pescara 0.995 Vibo Valentia0.808
Firenze 0.989Piacenza 0.613 Vicenza 0.993
Foggia 0.877Pisa 0.975 Viterbo 0.940
Forli -Cesena0.973Pistoia 0.954
Frosinone 0.998Pordenone 0.969
4In our view the relevance of the urban size and temporal dimension is
typified in [4,24]. [4:136] refer to the fact that even in underindustrial-
ized areas exhibiting slower growth, ultimately there will be enough
b
usiness to justify the establishment of a sheet metal shop, a truck
depot, etc., which eventually will result in a ‘critical mass of industrial
facilities’ which enables accelerated economic growth. [24:288] de-
scribes how the agglomeration of industries and firms is based on the
greater availability of goods and services related to the dominant Indus-
try. The author gives the example of transportation, a service estab-
lished based on the needs of a dominant industry, whose availability
attracts other industries to the area based on its lower cost compared to
other territories. These additional industries benefit also from the
community of business service suppliers in the area, e.g. legal, ac-
counting, printing, etc. Source: O wn elaboration on ISTAT da tabase (2001).

A. MARRA ET AL. 442
[9] describes this as: “But although cities may fairly be
called natural economic generators of diversity and
natural economic incubators of new enterprises, this
does not mean that cities automatically generate diver-
sity just by existing. They generate it because of the
various efficient economic pools of use that they form.
Wherever they fail to form such pools of use, they are
little better, if any, at generating diversity than small
settlements… At the other extreme, huge city settlements
of people exist without their presence generating any-
thing much except stagnation and, ultimately a fatal dis-
content with that place… Rather, something is wrong
with their districts; something is lacking to catalyze a
district population's ability to interact economically and
help form effective pools of use” [9:148].
Based on this, first we observe the evolution of H’ti
over the time span 1951-2001 in the largest provinces
and in the smallest ones (Figure 2a, b). In the first group,
18 provinces show high levels of entropy (above 0.5,
(a)
(b)
Figure 2. (a, b) Economic diversity in the 20 largest (a) and
20 smallest (b) provinces over the time period 1951-2001,
normalised values) in 1961, w
Source: ISTAT database (2001).
hile the process of in-
se
mallest provinces, the industrial pattern is less
im
creasing diversification i s much slower i n sm all provi nces.
Figure 3 (a) shows that in 1961 large provinces pre-
nt a more important economy and some basic Indus-
trial sectors (code and number of employees): retailing
(code = 16, n° employees = 749,978), constructions (19,
441,960), non electric machines, metallic carpen - try and
foundries (30, 369,459), wholesale trade (17, 227,673),
hotels and restaurants (4, 220,390), ma- chines and repair
plants (40, 198,791), clothing Indus- try (1, 181,978),
food and beverages (5, 164,830), electric machines for
communications (29, 152,939), non metal-bearing min-
erals (37, 151,677), chemicals (15, 146,464), metallurgy
(34, 137,098), communica- tions (18, 105,140), wood
industry (27, 104,823), shoes industry (12, 92,706),
credit (20, 87,497), other services to transport (9,
79,946).
In the s
portant and less diverse. Only a few sectors in this
(a)
(b)
Figure 3. (a, b)The economic pattern of the 20 largest (a)
tabase (2001).
and 20 smallest (b) provinces in 1961, Source: ISTAT da-

A. MARRA ET AL.443
group are shown to be relevant: retailing (16, 111,748),
constructions (19, 61,594), hotels and restaurants (4,
30,551), food and beverages (5, 25,286), machines and
repair plants (40, 19,306), clothing industry (1, 17,034),
wood industry (27, 16 ,947), wholesale trade (17, 16,765)
and non metal-bearing minerals (37 , 12,637).
Large provinces are also characterised by a strong
correlation between some blocks of economic activities
compared to small ones (see the graphical representation
of correlation matrices in Figure 4). We use the correla-
tion matrix of employment levels in each sector to better
identify similarities in terms of number of employees in
both sets of provinces. We obtain two 47 × 47 matrices
whose i,j entry is corr(empi, em pj), that is, the coefficient
of correlation between the employment level in the i- th
sector and the number of employees in the j-th sector.
Large provinces show a clear economic configuration
that is defined by a “recurrent” industrial pattern: aside
from hotels and restaurants and food and beverages,
there are retailing, wholesale trade, communication, con-
structions, credit and other services to transpor t, machine
(a)
(b)
Figure 4. (a, b)Correlation r the largest (a) and
smallest (b) provinces, Source: ISTAT database (2001).
sis in this paper investigates whether,
matrix fo
and repair plants are all correlated sectors (see red areas
in Figure 4). This result seems surprising since large
provinces are generally characterised by high levels of
entropy, often used in biology as a measure of disorder.
In contrast, small provinces’ industry structures, which
have low levels of entropy, conform less to a common
sectoral pattern.
Since correlation does not mean complementarity we
attempt some preliminary qualitative reasoning. A sub-
stantial difference between diversity meant as 1) the
source of knowledge spillovers, 2) basis for the
city-nursery and 3) source of a portfolio of economic
activities (which protect the relevant geographic area
from external shocks) is given by the degree of comple-
mentarity among industrial activities: the concept of the
city-nursery assumes that economic sectors are suffi-
ciently complementary to provide the basic goods and
services required for new entrepreneurial initiatives
while the portfolio theory is based on their diversifica-
tion and lack of interdependence. Moreover, although
knowledge spillovers may occur between very diverse
(i.e., not interrelated) industries, they are more likely in
sectors where people meet to discuss common problems.
The first two theoretical perspectives were discussed
in the introduction and in section 2. According to the
portfolio argument, economic diversity could be seen as
a strategy to spread risk, in order to protect a region from
external demand shocks (similar to a strategy of corpo-
rate diversification). High sectoral diversity in the re-
gional economy implies that a negative demand shock in
any of these sectors will have only moderately negative
effects on growth. On the other hand, a geographic area
specialising in one sector (or group of sectors with re-
lated demand) is at risk of serious slowdown in growth,
and high rates of unemployment [26].
Given the predominantly “instru mental” (or “facility”)
nature of the economic activities characterising the in-
dustry pattern in large provinces (including retailing and
wholesale trade, constructions and wood industry, non
electric machines, machines and repair plants and elec-
tric machines for communications, non metal-bearing
minerals and metallurgy, metallic carpentry and foun-
dries, chemicals, communications, credit and other ser-
vices to transport) we can assume a reasonable degree of
complementarity and, hence, argue i favour of the inter-
pretation of the city as nursery.
5. Conclusions
The economic analy
under which circumstances and to what extent diversity
of economic activities explains how Italian provinces
grow and evolve. We made use of a measure of entropy

A. MARRA ET AL. 444
between our proposed index (DDH ’i) and em-
pl
va
rked to boost growth in the largest provinces. In
19
s key-facilities
(e
e
969a) The Economy of Cities. Random
House, New York.
arch Policy, Vol. 11, No. 4, 1982, PP.
Dawn of History to the Present. Universi
iness Review, Vol. 38, No. 1,
ife. Random House, New York.
ban
f Products,”
in cities,” Journal of Political Economy,
9b, PP. 652-656.
e,
ine empirische Analyse
,” Re-
ly monopoly?” Mathematical Social
mic Review, Vol.
c Growth in the United
anufacturing Industry, Government
,”
ory,
litical Economy, Vol.
rt F. G. (2004). Urban
to account for two specific hypotheses: first, diverse in-
dustry patterns generate effects over the time; secondly,
the positive effects of economically diverse prov inces on
economic growth are amplified by the size of the prov-
ince.
We estimated the association per province and over
the timek
oyment levels. We found po sitive and v ery high results
for most provinces (64 provinces have a correlation co-
efficient higher than 0.90), while only 19 have a negative
and significant correlation coefficien t.
We also investigated empirically Jacobs’ assertion that
to understand cities we need to take direct account of the
Ame
[11]
rious efficient economic pools. Lacking information at
urban level, we made use of data collected at provincial
level.
We focused on the recurrent industrial p attern at 1961
that wo
61 large provinces hosted more important economies
and a more industrially diversified pattern compared to
small provinces which ha d less important and less diver-
sified industrial structures. Only a few sectors were im-
portant in small provinces. We find also that large prov-
inces are characterised by highly correlated blocks of
economic activities, despite large prov inces are generally
characterised by high levels of entropy (or disorder) . The
industrial structures in small prov inces, on the o ther hand,
tend not to conform to a common pattern.
Given that economic activities characterising the in-
dustrial pattern of large provinces play a
Scie
.g., retailing and wholesale trade, constructions and
wood industry, chemicals, communications, credit and
other services to transport) we showed that there is a
fairly high degree of complementarity supporting city-
nurseries, and to a lesser extent knowledge spillovers,
rather than the portfolio argumen t on economic activities
at local level.
6. Referenc
[1] Jacobs, J. (1
[2] F. M. Scherer, “Inter-Industry Technology Flows In The
United States” Rese
227-45.
[3] Bairoch, P. (1988) Cities and Economic Development:
from thety of
Vol. 23, No. 4, 1963, PP. 414-443.
[21] E. L. Glaeser, H. Kallal, J. Scheinkman and A. Shleifer,
“Growth In Cities,” Journal of Po
Chicago Press, Chicago.
[4] B. J. Chinitz, R. Vernon, “Changing Forces In Industrial
Location,” Harvard Bus
1960, PP. 126-136.
[5] Jacobs, J. (1984) Cities and the Wealth of Nations: Prin-
ciples of Economic L
[6] G. Duranton, D. Puga, “Diversity and Specialisation in
Cities: Why, Where and When Does it Matter?,” Ur
Studies, Vol. 37, No. 3, 2000, PP. 533-555.
[7] G. Duranton, D. Puga, “Nursery Cities: Urban Diversity,
Process Innovation, and the Life-Cycle o
rican Economic Review, Vol. 91, No. 5, 2001, PP.
1454-1477.
[8] J. V. Henderson, A. Kuncoro and M. Turner, “Industrial
development
Vol.103, No. 5, 1995, PP. 1067-1090.
[9] Jacobs, J. (1961) The Death and Life of Great American
Cities. Vintage Books, New York.
[10] J. Jacobs, “Strategies for Helping Cities,” American
Economic Review, Vol. 59,No. 4, 196
M. Attaran, “Industrial Diversity and Economic Per-
formance in U.S. Areas,” Annals of Regional Scienc
Vol. 20, No. 2, 1986, pp. 44-54.
[12] P. Haug, „Diversifikation und regionale Wirtschafts- und
Beschäftigungsentwicklung - E
für ausgewählte deutsche Gebiete“, Jahrbuch für
Regionalwissenschaft, Vol. 24, No. 2, 2004, PP. 177-195.
[13] R.W. Bostic, J.S. Gans and S. Stern, “Urban Productivity
and Factor Growth in the Late 19th Century,” Journal of
Urban Economics, Vol.4, No.1, 1997, PP. 38-55.
[14] T. Garcia-Mila, T.J. McGuire, “Industrial Mix as a Factor
in the Growth and Variability of States' Economies
gional Science and Urban Economics, Vol. 23, No. 6,
1993, PP. 731-48.
[15] W. B. Arthur, “Silicon Valley locational clusters: do in-
creasing returns imp
nces, Vol. 19, 1990, PP. 235-251.
[16] A. Ciccone, R.E. Hall, “Productivity and the Density of
Economic Activity,” American Econo
86, No.1, 1996, PP. 54-70.
[17] U. Lim, “Knowledge Externalities, Spatial Dependence,
and Metropolitan Economi
States,” Environmental and Planning A, Vol. 39, No.4,
2007, PP. 771-788.
[18] N. A. Phelps, “From Branch Plant Economies to Knowl-
edge Economies? M
Policy, and Economic Development in Britain's Old In-
dustrial Regions,” Environment and Planning C: Gov-
ernment and Policy, Vol.27, No. 4, 2009, PP. 574-592.
[19] C. Beaudry and A. Shiffauerova, “Who’s right, Marshall
or Jacobs? The Localization Versus Urbanization Debate
Research Policy , Vol. 38, No. 2, 2009, PP. 318-337.
[20] N. Rosenberg, “Technological Change in the Machine
Tool Industry, 1840-1910,” Journal of Economic Hist
100, No. 6, 1992, PP. 1126-1152.
[22] Hoover, E. M. (1948) The Location of Economic Activity.
McGrawHill, New York. Van Oo
Growth and Innovation. Spatially Bounded Externalities
in the Netherlands. Ashgate, Aldershot.
[23] Florence, S. P. (1953). The logic of British and American

A. MARRA ET AL. 445
conomic Review, Vol. Va-
industry. Routledge, London.
[24] B. J. Chinitz, “Contrasts in Agglomeration: New York
and Pittsburgh,” American E51, r
1961, PP. 279-89.
[25] Hoover E.M. and Vernon R. (1959) Anatomy of a Me-
tropolis. Cambridge, MA: Harvard University Press.
[26] K. Frenken, F.G. van Oort and T. Verburg, “Related
iety, Unrelated Variety and Regional Economic Growth,”
Regional Studies, Vol. 41, No. 5, 2007, PP. 685-697.