Company performance of Italian industrial districts

Company performance of Italian industrial districts
Master thesis Concept version
January 2010
Student: J.J. van ’t Hooft
Student: F.A. van der Wind
Supervisor: Dr. S. Phlippen
Supervisor: Prof. Dr. R. Boschma
Department of Applied Economics
Section of Economic Geography
Erasmus University Rotterdam
Utrecht University
1
Abstract
The existence of industrial districts forms a characteristic property of the Italian economy and has
significantly contributed to the economic development and prosperity of postwar Italy. Especially
since the 1970’s, industrial districts have been extensively studied by scholars from various
academic disciplines, creating a wide body of theoretical and descriptive literature. Attempts to
quantify distinctive aspects of industrial districts, however, remain limited.
This research uses company level financial accounting data to measure the ‘district effect’ by
comparing performance indicators of companies from the Italian footwear sector belonging to
industrial districts to those of isolated companies. In addition, the outcome of a series of in-depth
interviews held with entrepreneurs, senior managers, and experts will permit a qualitative
interpretation and enrichment of the quantitative analysis.
This research concludes that footwear producing companies located in Italian industrial district
outperform isolated companies on all three considered performance indicators. The observed
performance differential is attributed to a higher degree of specialization among industrial district
companies. Other factors considered by this research, however, do not distinctly explain observed
performance differences.
Keywords:
economic geography, industrial districts, industrial organization,
agglomeration economies, transaction costs economics
JEL Classification: G21, L20, L67, R12
Copyright © 2010 Jeroen van ‘t Hooft & Floor van der Wind
2
Copyright
The copyright of present thesis rests exclusively with the authors. No quotations from it may be
reproduced, published, altered, nor distributed without the authors’ explicit prior written consent.
Information derived from this thesis must be acknowledged.
Disclaimer
The information presented in this study is based on research carried out by the authors as part of
their master thesis at the Erasmus University Rotterdam and Utrecht University respectively. No
part of this thesis has been submitted elsewhere priory for any other degree program. All its
content lies with the authors, unless reference to others is made.
3
Content
Abstract ..................................................................................................................................................................... 2
Content ...................................................................................................................................................................... 4
1. Introduction ........................................................................................................................................................ 6
1.1 Problem definition and research question .................................................................................................................7
1.2 Focus ............................................................................................................................................................................................8
1.3 Scientific relevance ................................................................................................................................................................8
1.4 Societal relevance ..................................................................................................................................................................8
1.5 Outline .........................................................................................................................................................................................8
2. Research subject ............................................................................................................................................ 10
2.1 Particularities of the Italian economy....................................................................................................................... 10
2.2 Italian footwear sector ..................................................................................................................................................... 10
2.3 Fermano-Maceratese footwear industrial district .............................................................................................. 11
3. Theoretical framework................................................................................................................................ 14
3.1 Distinctive features of Italianate industrial districts ......................................................................................... 14
3.2 Typologies of industrial districts ................................................................................................................................ 14
3.3 Porter’s competitive advantages ................................................................................................................................. 22
3.4 Company performance measurement ....................................................................................................................... 24
3.5 Hypotheses and propositions ......................................................................................................................................... 24
4. Methodology .................................................................................................................................................... 36
4.1 Quantitative method.......................................................................................................................................................... 36
4.2 Qualitative method ............................................................................................................................................................. 42
5. Results................................................................................................................................................................ 48
5.1 Quantitative results............................................................................................................................................................ 48
5.2 Qualitative results ............................................................................................................................................................... 55
5.3 Summary ................................................................................................................................................................................. 62
4
6. Conclusion ........................................................................................................................................................ 64
6.1 Limitations ............................................................................................................................................................................. 65
6.2 Recommendations............................................................................................................................................................... 66
6.3 Final consideration............................................................................................................................................................. 66
7. Acknowledgements ....................................................................................................................................... 68
8. References ........................................................................................................................................................ 69
8.1 Literature ................................................................................................................................................................................ 69
8.2 Websites................................................................................................................................................................................... 76
9. Appendices ....................................................................................................................................................... 77
9.1 Questionnaire footwear companies ............................................................................................................................ 77
9.2 Overview tables and figures ........................................................................................................................................... 80
5
1. Introduction
Since early days, the location of economic activity across the geographical space has welcomed a
profound interest from researchers of different academic disciplines (Von Thünen, 1826; Weber,
1909; Christaller, 1933). Recently, the field of economic geography has regained particular
attention, especially through the work of 2008 Nobel Prize laureate Paul Krugman, which unifies
the previously disparate research fields of international trade and economic geography (Krugman,
1991).
Within the field of economic geography, agglomeration economies have been thoroughly studied
(Marshall, 1920; Jacobs, 1969; Rosenthal & Strange, 2003). Positive agglomeration economies refer
to various sorts of savings or benefits derived from the clustering of activities external to the
company and are kinds of positive externalities. An example of clustering of economic activities
takes place in so called industrial districts (Becattini, 1990), which are the main interest of present
research.
The economic miracle of Italy in the 1970’s, during which small and medium sized enterprises
(SMEs) started contributing substantially to the economic development and welfare in Italy, has
initiated a large stream of theoretical work focusing on industrial districts (Becattini, 1990; Sforzi,
2002). The concept of industrial district was first introduced by Alfred Marshall (1920) and has
enjoyed a revival from the 1970’s due to the work of Giacomo Becattini and his Florentine School.
Marshall names three reasons for companies to agglomerate: (1) knowledge spillovers, (2)
decrease of transaction costs, and (3) easier access to skilled labor.
All three factors are briefly reviewed here. Knowledge spillovers occur in different ways. For
example, tacit knowledge1 needs frequent face-to face interaction in order for the knowledge
getting widespread. Collocation fosters this sort of interaction (Karlsson et al, 2004). The second
reason for clustering consists of companies enjoying facilitated access to specialized products.
Because of the agglomeration of specialized companies, a larger market for specialized products in
the district is created, which decreases transaction costs (McCann, 2004). The third reason reflects
a facilitated access to specialized personnel: because of the agglomeration of companies, the district
itself can act as a magnet, attracting skilled laborers (McCann, 2004). The role played by highly
specialized labor, or a unique mix of skill assets, is considered highly important in stimulating the
district (Wolfe, 2001).
The combination of knowledge spillovers, decreasing transaction costs, and skilled labor create
competitive advantages of being located in an industrial district. Becattini (1990) concludes that
many industrial sectors in Italy possess these phenomena. An Italian industrial district (hereafter
referred to as IID) is defined by Becattini as:
1
Tacit knowledge: Information or abilities that can be passed between persons with personal contact, but which
cannot be passed on in a written form (Collins, 2001 p. 72).
6
‘A socio–territorial entity characterized by the active co-presence, in a territorially circumscribed,
naturally, and historically determined area, of a persons’ community and of an industrial enterprises’
population. In the district, at the opposite of what happens in other environments (for example, in the
manufacture city), community and enterprises tend, say, to mutually interpenetrate.’
The difference between IIDs and Marshallian industrial districts is that in Becattini’s notion of IIDs
all local actors (e.g. shoe producers, subcontractors, governmental organizations, and financial
intermediaries) consciously cooperate and that all local actors are strongly rooted in the local
territory (Alberti, 2003). IIDs are formed by the nature of the internal relationships between
community and its companies. These relationships are extensively explored by several authors
(Brusco, 1990; Pietrobelli, 1998; Signorini, 2000; Paniccia, 2002; Boschma & Ter Wal, 2005).
The potential benefits of the agglomeration of companies in IIDs are believed to lead to superior
performance and higher income for the inhabitants of such areas (Becattini, 1990). However, recent
authors have started questioning the superior performance of IIDs. Hadjimichalis (2006), for
example, argues that IIDs’ performance depends on different factors, which are not fully integrated
in the current theory on IIDs. An example of such a factor is the ignorance of (unpaid and
unregistered) labor by family members. Therefore, the factual performance of companies active in
IIDs and the origin of that performance are unclear (Hadjimichalis, 2006; Rabellotti et al, 2009).
Despite the large existing body of literature concerning IIDs, there is not much empirical evidence
related to explanations for performance of companies located in IIDs. Porter’s (2008) five forces
framework forms a holistic tool to explain differences in profitability among companies and is
applied to assess company performance in this context. The position of the company in relation to
new entrants, substitutes, customers, suppliers, and competitors can be used when explaining
possible profitability differentials between companies located inside IIDs and outside IIDs.
1.1 Problem definition and research question
This study tries to provide an answer to the question how the performance of companies in IIDs
relates to its non-district counterparts. In addition, it explores factors that contribute to the
performance of companies active in IIDs. This leads to the following research question:
Do companies located in industrial districts outperform their non-district counterparts and why?
In order to eventually answer the main research question, this paper uses two sub-questions. The
sub-questions are formulated such that they lead to the answer of the main research question and,
moreover, provide additional insights.
1. How do IID companies perform compared to their non-IID counterparts?
2. What are possible reasons for potential observed performance differences?
7
1.2 Focus
IID companies are those companies located in IIDs and further referred to as IDCs. Based on ISTAT2
2001 census data, 156 IIDs are identified (Rabellotti, 2009). In order to make this study feasible, it
investigates the case of one particular agglomeration through a series of in-depth interviews held
with entrepreneurs, senior managers, and experts from the so called Fermano-Maceratese footwear
district, located in the Marche region in central Italy. This study draws a comparison between
companies located within the mentioned footwear agglomeration and other Italian footwear
companies located in more isolated locations.
1.3 Scientific relevance
Attempts to offer a measurement of the performance of IIDs on a comparative basis are still few
(Paniccia, 1999). By assessing the performance of companies located in IIDs compared to
companies located outside IIDs, this study contributes to the creation of such an empirical
measurement. By investigating factors contributing to the performance of IID companies, this study
contributes to the discussion of possible explanatory factors of mechanisms responsible for IIDs
company performance.
In addition, this study includes the discussion of two ‘hot topics’ in economics, which have attracted
particular attention recently: (1) economic geography (Nobel Prize 2008, Paul Krugman) and (2)
transaction costs (Nobel Prize 2009, Oscar Williamson).
1.4 Societal relevance
The discussion about the performance industrial districts is relevant to a wide group of
stakeholders including policy makers, entrepreneurs, managers, and employees. So far, industrial
districts are supposed being a beneficial policy tool for creating regional development and welfare.
IIDs would serve as a ‘best practice’ model for stimulating local development. However, recent
studies indicate that the configuration of IIDs is not applicable to every single case and reality
(Hadjimichalis, 2006).
The results of present study can contribute to the understanding of the origin of company
performance in IIDs and its impact on the local area. This is important to enable policy makers
developing and implementing effective policy instruments and policy measures.
Since this study takes the Italian footwear sector as its focus of analysis, it may be of interest to
entrepreneurs, managers, and employees active in the Italian footwear sector as well. Given the
specific characteristics of this sector, present study may give insight into particularities of the
footwear sector and can support persons involved in this sector in their strategic decision making.
1.5 Outline
The next chapter introduces the research subject and describes the history and present situation of
IIDs. Chapter 3 presents the theoretical considerations framing this study and leads to a number of
hypotheses and propositions. Subsequently, the methods used for obtaining and elaborating both
2
Istituto Nazionale di Statistica, Italian Bureau of Statistics
8
quantitative and qualitative data used in the empirical analysis are introduced and discussed in
Chapter 4. Chapter 5 presents and interprets the results of the empirical and qualitative analyses.
Finally, Chapter 6 will present the conclusion of this thesis by answering the main research
question, discussing the limitations of this research, and suggesting directions for further research.
9
2. Research subject
This chapter starts with a brief overview of the Italian economy and describes a number of its
particular characteristics. After this, the development of the Italian footwear sector is presented.
Subsequently, a description of the history and current situation of the IID studied for the purpose of
this research, the Fermano-Maceratese footwear district, is provided. This description is completed
by a presentation of the different actors active in this district.
2.1 Particularities of the Italian economy
Since the end of the Second World War, the Italian economy has transformed itself from a merely
agriculturally based economy into a highly industrialized one: nowadays, the Italian economy
strongly depends on traditional manufacturing industries (for example clothing, furniture, and
shoes and leatherwear) (Gola & Mori, 2000), characterized by relatively low levels of capital
intensity and only moderate use of advanced technologies compared to Western-European
countries. These traditional manufacturing industries are moreover characterized by relatively low
entry barriers and by a relatively elastic consumer demand (Bronzini, 2001).
The overall Italian economy is characterized by a large presence of small and medium sized
enterprises (hereafter SMEs): the share of companies with less than 250 employees is 70%, which
is the highest among industrialized countries (Brusco & Paba, 1997). Remarkably, these SMEs are
predominantly owned and managed by single families. Another particular aspect of the Italian
economy is the substantial economic disparity between the more developed northern and central
regions on the one hand and the less developed southern regions (including the islands Sardinia
and Sicily) on the other hand (Signorini, 2000).
Together with the strong presence of family owned SMEs, the Italian economy exhibits a strong
presence of industrial districts. Almost one fourth of the total Italian population lives in areas
defined as ID, while among one third of the Italian population works within IIDs (ISTAT census
data, 1996).
2.2 Italian footwear sector
The footwear sector started to expand in the 1960’s until mid 1980’s, exploiting its relative labor
costs advantage compared to other European countries. In this period, the sector became more
export orientated and Italy developed itself as the leading footwear exporting country of Europe
and second largest footwear exporter of the world (Rabellotti, 1995). During the 1980’s and the
beginning of the 1990’s, this position was challenged by growing globalization resulting in
increased competitive pressure from low labor cost countries, predominantly located in Asia and, to
a lesser extent, Eastern Europe. Italian footwear companies reacted differently to this threat: some
of them started to aim at higher market segments, competing on dimensions like quality, design,
brand, and marketing rather than on price. Other footwear producers adopted a strategy of
reducing production costs by (partially) outsourcing relatively labor intensive activities of the
production process to Eastern Europe or Asia (Rabellotti, 2003).
About 13% of the IIDs established by ISTAT operate in the footwear sector (Rabellotti, 2009). The
10
footwear sector is mainly concent
concentrated
rated in IDs: 66% of total employment is situated in IIDs (De
Blasio & Di Addario, 2005). The footwear sector is concentrated predominantly in IDs located in the
Veneto, Toscana, Marche, Campania, and Puglia regions (Rabellotti, 2003), as shown in Figure 2.1.
Figure 2.1Main footwear producing agglomerations in Italy
2.3 Fermano-Maceratese
Maceratese footwear industrial district
The footwear producing Fermano
Fermano-Maceratese district (named after the provincial towns Fermo and
Macerata) is located in the Marche region in central Italy. The origin of the district lies in shoe
masters arriving as early as 1870 in this area. The footwear sector in that area was struck by a crisis
in the beginning of 1900, after which the production was reorganized and small workshops
specializing
alizing in various phases of production started to arise. This organizational innovation formed
the basis for the exponential growth of the district during the 1960’s (Aureli et al, 2008). The
district was more cost-efficient,
efficient, because the production specia
specialization
lization allowed benefiting from
scale economies, a particular feature of industrial districts still believed to hold nowadays. During
the 1980’s, facing growing global competition, some companies started outsourcing phases of the
production chain outside the district to low labor costs countries and began investing in branding
and marketing activities. Currently, the Fermano
Fermano-Maceratese
Maceratese district is the largest district in Italy,
housing around 4’000 companies active in footwear producing and related activit
activities
ies (Aureli et al,
2008). The district produces women, men, and children footwear. The district mainly focuses on the
medium-upper
upper level market segment (Rabellotti, 2003). Figure 2.2 gives a general presentation of
the actors actively involved in the Ferma
Fermano-Maceratese footwear district.
11
Figure 2.2 Actors involved in the Fermano
Fermano-Maceratese footwear district
Shoe producers represent medium
medium-sized
sized or large companies, which adopted a corporate strategy
focusing on branding and marketing. They have both national and (increasingly) international
markets for their products. Part of their production is outsourced to subcontractors inside the
district.
Subcontractors represent companies, which are highly specialized in a single phase of the
production chain. Examples of these are tomaifici (specialized in uppers), tacchifici (specialized in
heels), and suolifici (specialized in sol
soles).
es). Besides these specialized subcontractors, other forms of
subcontracting exist (calzaturifici
calzaturifici di assemblaggio) in which relatively small companies assemble
components into final products commissioned by larger shoe producers.
Financial intermediaries consist
onsist of local, regional, and national banks. An example of a local bank is
Banca di Credito Cooperativo, which is a nationwide cooperative bank with multiple local
subsidiaries throughout Italy. Banca delle Marche is a regional bank, which is only present
presen and
active within the regional territory and moreover primarily invests its funds in the Marche region
itself. Finally, UniCredit Bank is a large conglomerate financial services provider active even beyond
national borders.
The Italian public sector has four different hierarchical administrative layers: (1) comuni
(municipalities), (2) provincie (provinces), (3) regioni (regions), and (4) governo nazionale (national
government).
ment). For the footwear sector regional governments play an important role, since Italy’s
It
industrial policy is predominantly established and executed at the regional level.
Apart from the abovementioned governmental organizations
organizations,, there exist as well several trade
associations related to the footwear sector
sector. They are formed by its members from the footwear
sector and represent their members’ interest. ANCI stands for the Associazione Nazionale
Calzaturifici Italiani and is the Italian national footwear association. They represent over 1’000
predominantly larger companies and their services consist of the organization of promotion
12
activities and lobbying at national level for the interests of the sector.3 CNA is the National
Confederation of Artisanship and Small and Medium-Sized Enterprises. The organization has more
than 600’000 associated members and aims at strengthening and developing the crafts of the
medium and small enterprise sector in Italy.4 Confindustria represents 135’000 companies of all
sizes in the manufacturing and service industries in Italy. Their activities consist of lobbying
activities aimed at encouraging improved competitiveness for Italian enterprises represented by
regional affiliates.5 Confartigianto represents 520’000 artisan companies (micro, small, and medium
sized) and autonomous entrepreneurs from all industrial sectors with the purpose of increasing the
added value of its associated members, lobbying activities, establishing and improving economic
relations and relations with labor unions, institutions, and other associations.6
3
www.ancionline.com
www.cna.it/eng
5
www.confindustria.it
6
www.confartigianato.it
4
13
3. Theoretical framework
This chapter presents the different theoretical aspects that frame the underlying base of this study.
The subsequent paragraph presents the ten distinctive features of IIDs as established by the
‘godfather’ of the theoretical body on Italian industrial districts, Giacomo Becattini (paragraph 3.1).
Paragraph 3.2 discusses different contributions by other authors from the industrial district
literature, proposing different typologies for IDs. The features presented in paragraph 3.1 are
related to Porter’s framework of five competitive forces in paragraph 3.3 Paragraph 3.4 addresses
company performance indicators and reviews a number of studies on company performance
measurement. This chapter’s final paragraph leads to a series of propositions and hypotheses,
which are analysed throughout Chapters 4 and 5.
3.1 Distinctive features of Italianate industrial districts
Becattini (1990) identifies ten specific features that characterize Italian industrial districts. Some of
them originate from the concept of industrial districts as proposed by Marshall (1920).
1. Local community of people as a social feature: the local community of an IID is characterized
by homogeneity in values and believes. These values need to have a nature, which is
beneficial for the survival of the district (for example, encouragement of entrepreneurship).
Values are spread through IIDs by institutions such as the family, company, school, local
authorities, labor unions, trade associations, and political parties. Due to the intangible
characteristic of most of these institutional factors, researchers find it difficult to measure
and assess their impact (Omiccioli, 2000).
Furthermore, the social feature comprises as well a ‘natural resistance’ towards values from
outside and the use of ‘double standards’: one standard towards members of the local
community (insiders) and another standard towards ‘strangers’ (outsiders).
2. Population of companies: companies inside districts belong mainly to the same industrial
branch and are usually specialized in just one or few phases of the production chain. This
enables specialization, which is assumed to enhance performance. The concept of
‘industrial branch’ should be broadly interpreted in this context: next to the principal
industry, ‘auxiliary industries’ (compare Figure 2.1) are comprised as well.
Another characteristic of companies located in industrial districts relates to its products:
technically suitable products for a district are fairly differentiated products whose market
demand is inconstant over time.
Finally, the presence of pronounced and strong inter-firm relationships – mainly via kinship
and friendship between actors in the district – is a remarkable feature, unique to the nature
of IIDs.
3. Human resources: this feature relates to a pool of specialized labor present inside the
district. Within the district, every actor tries to perform activities which best suit their
capabilities and ambitions. The tendency of the district to optimally reallocate its human
resources forms one of the conditions for the districts’ competiveness (Alberti, 2003).
14
Specialized companies in the district attract and retain the most skilled workers. If workers
change employer, but remain active within the district, the worker’s specialized knowledge
and accumulated experience still remains part of the district and is not lost. Therefore,
specialized knowledge is considered a public (that is, non-rival and non-excludable) good
inside the IID, freely available to all its members.
Next to the availability of a specialized labor pool, the high degree of entrepreneurship is
considered a characteristic feature related to the human resources of IIDs. In the local
culture of many IIDs, being self-employed means achievement and enhances social status.
Becattini (1990) proposes the concept of the ‘pure entrepreneur’, a figure ‘translating all
capabilities, which are latent in the historical heritage of the district into products that can be
sold in the market’.
Finally, the presence of a large base of lavoratori a domicilio (home workers) is regarded
specific for the human resources of IIDs and forms the link between the IID’s flexible
productive system on the hand and its social system via its families on the other hand.
4. Market: the market within an IID is in line with the distinctive image of its product’s
characteristics (related to a certain quality standard and accessory conditions of the
transactions). The purchase of raw materials attracts a substantial number of specialized
buyers, allowing for economies of scale to emerge.
5. Competition and cooperation: the balance between competition and cooperation among
companies operating inside the district is a fifth distinctive feature of IIDs. This balance
translates itself into a high degree of vertical cooperation, in which companies seek
economies of scale in complementary activities and a high degree of horizontal competition
between companies performing similar activities (substitutes).
From a social point of view, competition inside IIDs does not derive from the strive for
driving out rivals at the first place, but rather from (1) the desire of improving the
(economic and social) position of IIDs’ entrepreneurs and workers and their families and
(2) the continuous motivation of IIDs’ entrepreneurs and workers to discover economic
solutions superior to current common standards available outside IIDs.
A final characteristic related to competition and cooperation is presented by the
establishment of ‘local’ price levels for goods and services frequently traded inside IIDs.
Even though these local prices reflect to a certain extent the (internationally) widely
prevailing prices -the latter being a simple result of regular demand and supply conditionslocal prices tend to fluctuate less and, as a consequence, offer IIDs’ actors considerable
certainty of stable income and production costs levels.
6. Adaptive system of production organization: the various interests of IIDs’ different actors
adapt quickly to the changing organization of its production system, which increases the
efficiency of each single production phase. This adaptive system optimizes the governance
structure and industrial organization of IIDs by minimizing transaction costs, largely caused
by the presence of a mechanism of ‘social control’.
15
7. Technological change: rather than the simple outcome of a hierarchical decision making
process (which may cause natural resistance as it is assumed impairing the value of human
capital), the introduction of technological change in IIDs results from a gradual social
process of self-awareness, present among all its different actors.
Therefore, the introduction of technological change in IIDs is supposed to be widely
supported by the members of its community. This is crucial for a production system which
is relatively more intensive in labor than in capital. This open attitude towards
technological change has a positive effect on the ability to pick up external knowledge. This
ability is defined by a company’s absorptive capacity (Cohen & Levinthal, 1990).
Absorptive capacity is largely a function of prior related knowledge, which relates to the
ability to recognize value, assimilate, and apply new information to commercial purposes.
The prior related knowledge is usually measured in terms of employees’ educational
background within organizations and the investments in R&D activities of a company
(Cohen & Levinthal, 1990). Small companies usually lack sufficient financial resources for
R&D investments and as a result they generally not possess a distinctive R&D budget
(Boschma & Ter Wal, 2005). It is therefore difficult to assess the absorptive capacity of
small IDCs.
8. Local credit system: a disadvantage of small companies compared to large companies is the
limited access to external credit of the first compared to the latter, acknowledging for the
crucial importance of credit access for economic development to be possible. Inside IIDs
this disadvantage is resolved to some extent by the presence of ‘local banks’.
Local banks emerged and established within the local community and are closely linked to
local entrepreneurs. The roots of the local bank inside its own territory allow the bank to
better value the personal qualities of entrepreneurs and to better evaluate the prospects of
a proposed investment project.
The necessary condition for this system to work successfully is the absence of moral hazard
and opportunistic behavior of those persons requesting credit. Opportunism in this context
is defined as ‘an effort to realize individual gains through a lack of candor or honesty in
transactions’ (Williamson et al, 1975). This is plausible for IIDs given the presence of
mechanisms of social control present in the local community as mentioned under features 1
and 6.
9. Sources of dynamism: this feature consists of the continuous decision of industrial district
companies to execute activities inside or outside the company followed by the concern to
outsource activities inside the district or outside the district. This decision is guided by both
(1) economic as well as (2) cultural considerations.
The economic consideration mainly concerns a cost-benefit trade-off with respect to the
make or buy decision. One should note, however, that the concept of cost within the context
of IIDs does not depend on purely economic terms, but on the historical and cultural factors
defining the IID’s identity. The cultural consideration, on its turn, is shaped by the values
and norms of the district among others.
16
10. Consciousness, class, and locality: this feature of IIDs is translated by the feeling of local
actors belonging to the local community. It relates to production organization in two forms:
(1) endogenous relationships (between local culture, local community, and local economy)
and exogenous relationships (effects on the district of external markets, entire society, and
culture).
These specific features have an impact on company performance inside the district and the
performance of the district as a whole. It is referred to by some authors as the ‘canonical7 model’ of
IIDs (Alberti, 2003). The features characterizing IIDs are explored, assessed, and adjusted by
numerous researchers.
Pietrobelli (1998), for example, analyzes local factors, which explain IIDs’ competiveness by means
of a quantitative method, namely by measuring the export propensity. The measurement of the
factors is difficult because of their intangible nature. Pietrobelli uses six proxies for the factors
considered. Table 3.1 presents those factors explored by means of proxies and their relationship
with the features of Becattini et al (1990).
Feature in
Becattini
Factor in Pietrobelli
Proxy
human skills
number of educated people in the community
human skills
family-employed workers/active population
human skills
index of institutional performance
entrepreneurial culture
entrepreneurs in industry/active population
availability of financial
employed in financial services/active population
services
5
tradition of intercompany size and number of companies
company cooperation
Table 3.1 Explaining factors and proxies
Source: Becattini (1990) and Pietrobelli (1998)
3
1
1
1
8
The index of institutional performance is computed by Putnam (1993) in an earlier study. The data
concerning the other factors are derived from the 1991 Italian industry and population census
(ISTAT, 1991). Pietrobelli uses a sample of 37 industrial districts. The analysis shows that the
availability of human skills and financial services are positively correlated to IID performance. The
degree of secondary education related to technological skills of the population results the most
important, rather than the degree of university education. The positive correlation of the
availability of financial services could be explained by the expectation that local agents better know
their clients, enabling them to customize their services and reducing uncertainty. Family tradition
is negatively related to IID performance. Pietrobelli suggests that this is the result of an incorrect
proxy or a new development, in which the family only holds managerial positions in the company.
The remaining three factors are not correlated to IID performance. The study shows that not all
features of Becattini’s IID contribute to IID performance and that more statistical research is
required to be able to empirically test the hypotheses supposed by theoretical research.
7
canonical: from Greek κάνον, reduced to the simplest and most significant form possible without losing generality
17
3.2 Typologies of industrial districts
Throughout the years, different scholars have presented different typologies for industrial districts.
This paragraph briefly reviews three contributions from prominent scholars, providing an
overview of currently existing views on IIDs.
3.2.1 Four IDs typologies according Markusen (1996)
Markusen (1996) distinguishes four typologies of industrial districts, namely (1) Marshallian
industrial districts, (2) hub-and-spoke districts, (3) satellite industrial platforms, and (4) state
anchored industrial districts.
Table 3.2 below outlines the hypothesized features of the first typology extending for features
specific for Italianate industrial districts. Since types (2), (3), and (4) are not particularly relevant to
present research, they are omitted in the table.
Feature
Marshallian industrial districts
Italianate variant
(in addition to features aside)
Business structure
dominated by small,
locally owned companies
relatively low
highly flexible internal
to the district
high degree of family ownership
workers committed to district
rather than to companies
low with companies
external to the district
long-term local contracts
and commitments
substantial among
buyers and suppliers
disproportionate shares of workers
engaged in design and innovation
high among competitors to share risk
and to stabilize market
transactions based on trust rather than
on formal written contracts
high among highly specialized
producers of components
strong local government role in
regulating and promoting core industries
Scale economies
Labor and personnel
Cooperation and
competition
Local transactions
Intra-district trade
Institutions
Financing
specialized sources of finance, technical
expertise, business services available in
district outside of companies
exchange of personnel between
customers and suppliers
strong trade associations providing
shared infrastructure-management,
training, marketing, technical and
financial support, mechanisms for risk
sharing and stabilization
presence of local banks rooted in local
territory
key investments made locally
Table 3.2 Features of Marshallian and Italianate industrial districts
Source: based on Markusen (1996)
One should note that, compared to the classical Marshallian concept of industrial districts, the
Italianate variant is unique in its presence and important role attributed to institutions, which are
completely absent in the Marshallian notion. Moreover, one can note a higher degree of general
18
informatlity with Italiante IDs (importante role of family, absence of written contracts, exchange of
personnel amongst workers and suppliers).
3.2.2 Four IIDs typologies according Brusco (1990)
In his 1986 publication, Brusco presents three different typologies of IIDs. This classification is
expanded by a fourth one in Brusco (1990). His taxonomy is based on the historical evolution of
IIDs from the end of the Second World War until the early 1980’s and comprises (1) traditional
artisan model, (2) dependent subcontractor model, (3) Mark I model, and (4) Mark II model. Table 3.3
summarizes the main characteristics of Brusco’s classification.
Traditional
artisan
Emergence
Key
characteristics
Dependent
subcontractor
1950’s - 1960's
end 1960’s
dualism between
northern industrial
and southern
artisan companies
dualism between
large and small
companies
Mark I
Mark II
mid 1970’s
interaction between
companies and
institutions
early 1980’s
emergence of
‘real services’ (collective
in-kind services for IDCs
offered by institutions)
emergence of
auxiliary industry
Skill
accumulation
training on the job
equal both in
small and large
companies
increased by interfirm interaction
hindered by inertia of
district as a whole
Wage levels
lower than average
lower than average
average
Market
monopolistic
competition
vertical
oligopsony
slightly lower
than average
vertical
cooperation
Typical products
customized
products,
small batches
horizontal perfect
competition
specialized
intermediate
components
horizontal perfect
competition
final goods,
components,
auxiliary goods
horizontal monopolistic
competition
-
Typical
industries
clothing,
carpenters
-
textile, furniture,
footwear
-
Technology
labor intensive
low technology
labor intensive,
low technology
similar technology
for both small and
large companies
transition to higher
technological standards
vertical cooperation
Table 3.3 Four IIDs typologies
Source: Brusco (1990)
The Mark I typology of industrial districts explicitly mentions footwear as one of the typical
industries taking place. Next to that, another interesting observation can be made: whereas the first
two typologies mainly differ on issues related to disparities between northern industrial companies
19
and southern SMEs, for the last two typologies the role played by institutions is considered as
important. The difference between the respective typologies with respect to wage levels is
remarkable as the wage levels of IIDs employees change over time from below average in the
1950’s to average in the 1980’s. This raises questions about nowadays’ IID wage levels compared to
the national average.
3.2.3 Four IIDs typologies Paniccia (1999)
Paniccia (1999) analyses 1991 data concerning 39 Italian industrial districts. She explores 31
independent variables characterizing IIDs. The variables which mostly explain the observed
variance between districts are used to identify four different typologies of IIDs, namely, (1) craftbased, (2) canonical, (3) concentrated, and (4) embryonic.
The distinctive features of the craft-based districts consist of the relatively large proportion of
entrepreneurs and self-employed workers, a low rate of schooling drop out, extensive families, high
rates of unionization, and political subcultures. The footwear districts of Vigevano (located in
northern Italy (see Figure 2.1), originally footwear production, nowadays predominantly machinery
for the footwear producing industry) and Fermano-Maceratese belong to this type of IIDs.
‘Canonical’ districts are those IIDs, which exhibit the largest number of Becattini’s ‘canonical’
features. Those comprise a high degree of specialization, a high proportion of self-employed people
and auxiliary workers, a wide extension of the leading industry, and entrepreneurial density. Few
districts fall into this second typology.
The third typology, concentrated districts, is characterized by a high proportion of office workers
(‘white collar workers’), a higher degree of larger companies with more than 50 employees, and
limited division of labor.
The fourth industrial district typology includes IIDs characterized by low levels of specialization
and manufacturing industry. They exhibit a high rate of unemployment and illiteracy, which
represents a poor social performance. Most IIDs located in southern Italy are considered
embryonic.
Paniccia (1999) proposes two additional typologies to cover observed IID properties not covered
yet in the notion of Becattini’s IIDs’ dichotomy ‘canonical’ versus ‘non-canonical’. Her study reveals
as well that her four different typologies of IIDs are characterized by performance differentials.
Concentrated IDs have the best overall performance, which proofs that canonical IIDs are not the
only ones performing well and that other configurations are suitable for wealth creation as well.
Moreover, the analysis shows that over time the number of canonical IIDs declined, indicating the
rise of other configurations.
Summarizing, the ‘canonical’ IID of Becattini has specific features; some of which have been
assessed in other studies. Pietrobelli (1998) confirms the beneficial impact of financial
intermediaries and the presence of human skills on IIDs, however, at the same time raises questions
about the benefits caused by institutions and inter-company cooperation. Paniccia identifies four
different typologies of IIDs and marks footwear districts as craft-based IIDs. Both studies
(Pietrobelli, 1998; Paniccia, 1999) provide evidence for the existence of other successful
configurations and form a base for further research. Some aspects of canonical IIDs (labor market,
degree of specialization, competition and cooperation, institutions, and financial intermediaries)
20
therefore deserve further attention in order to assess their impact on company performance of IIDs.
Brusco (1990) and Pannicia (1999) both indentify four different typologies of IIDs and provide
further research together with Pietrobelli (1998). Some features (institutions, market, and labor
market) appear to be dynamic over time as different configurations of IIDs develop. It is therefore
interesting to see what the current situation and impact of these features in IIDs is.
So far, this chapter presented an overview of the ten characteristic IID features as proposed by
Becattini (1990) and various industrial district typologies from recent theories on industrial
districts (Markusen, 1996; Brusco, 1990; Paniccia, 1999). Among the features presented in
Becattini (1990), different features of IIDs are related to IIDs’ company performance. Four of these
features are elaborated in more detail throughout the remainder of this study. Among the ten
distinguishing features of Becattini (1990), technological change and absorptive capacity are not
considered further, given the intangibility of these features and related measurement difficulties.
Moreover, the role of family ownership is not elaborated on further, as family ownership is a
common phenomenon in entire Italy, and not for IDCs in particular.
The next paragraph introduces Porter’s model, presenting five forces affecting company
performance.
21
3.3 Porter’s competitive advantages
Porter describes that all companies in whatever industry are influenced by five competitive forces.
Differences in company profitability are explained by different strategic behavior. The five forces
shape the nature of the competition in an industry. Differences between companies’ profit potential
can be determined by means of the collective strength of Porter’s five forces (Porter, 1979, in
Faulkner, 2002). The forces are presented in Figure 3.1 and consist of: (1) threats of new entrants,
entrants
(2) bargaining power of new customers
customers, (3) bargaining power of suppliers, (4) threat of substitute
products, and (5) competition among current competitors
competitors.
Figure 3.1 Porter’s five competitive forces framework
Source: Porter (2008)
3.3.1 New entrants
Porter (1979, in Faulkner, 2002) describes that new entrants bring new capacity and present a
threat to the existing market shares of incumbent companies. The significance of the threat of new
entrants depends on the entry barriers of the industry. Porter discusses several entry barriers:
barriers
• economies of scale:: cost advantage, derived from economies of scale, for the existing
companies demands a large investment of the new entrant to be acquired;
• product differentiation:: power of brands of existing companies implies a large investment in
advertising
vertising by new entrants to overcome customer loyalty;
• financial capital requirements
requirements:: large financial investments necessary for up-front
up
advertising, R&D, and fixed assets, and to overcome start
start-up losses;
• cost disadvantages independent from size
size: these advantages of existing competitors originate
from multiple factors such as a favorable location, access to raw materials, government
subsidies, and the experience of existing companies;
• access to distribution channels
channels: degree to which it is possible for new
w entrants to use existing
distribution channels;
22
•
governmental policy: governments can limit entry by their control on license requirements,
safety, and pollution regulations.
Apart from entry barriers, potential expectations of entrants about the reactions of existing
companies can also influence the decision to enter an industry (strategic dynamic behavior).
3.3.2 Suppliers
Companies need supplies to be able to operate. These supplies can consist of raw materials,
knowledge, and/or labor. The bargaining power of suppliers of these resources has a large
influence on the industry’s profitability. A substantial bargaining power enables suppliers to
capture more of the added value. Bargaining power is increased if:
• there are less suppliers than buyers;
• suppliers are less dependent on buyers for their revenues;
• buyers face substantial switching costs;
• suppliers offer specialized products. For example, employers possessing specific skills,
which are irreplaceable for the production of the goods;
• there is no substitute for the delivered supplies;
• there is a credible threat for the buyers to be vertically integrated by their suppliers.
3.3.3 Buyers
Bargaining power of buyers or customers represents the reverse of a strong suppliers’ bargaining
power. Customers can capture value by demanding higher quality or lower prices. The factors that
determine customers’ bargaining power are similar to the factors that determine suppliers’
bargaining power. However, the bargaining power of customers increases even further if customers
are price sensitive. Price sensitivity is increased if:
• the product bought by the buyers represents a large fraction of its costs. If this fraction is
large buyers are more tended to bargain for it;
• the buyer group earns low profits or is pressured to cut down purchasing costs;
• the quality of the goods delivered by the buyers is not heavily influenced by the supplies of
the industry;
• the goods of the industry have little effect on the buyers’ other costs.
Assemblers and distributors represent intermediate customers; they gain additional bargaining
power if they can influence the purchasing decisions of the customers downstream. This can be
overcome by producers through exclusive contracts with distributors or by marketing directly
towards final customers. Producers of components try to gain bargaining power over their
intermediate customers by creating preferences for their components from downstream customers.
3.3.4 Substitutes
Substitutes are able to perform a similar function as the industry’s product in a different manner.
Substitutes generate a maximum for prices of the products in an industry. The threat of substitutes
increases if switching costs for the buyers are low. The threat can also increase if competition in the
industries for substitutes is fierce, leading to price reductions or better product performance of the
substitutes.
23
3.3.5 Rivalry among existing competitors
Existing competitors all try to gain or at least sustain their market shares. Rivalry is among others
represented by price discounting, introduction of new products, and advertising campaigns. Profit
margins are influenced by the intensity of rivalry. Porter indicates that the intensity of rivalry
increases if:
• there are numerous existing competitors that are equal in market power;
• industry growth is slow, which increases the fight for market share;
• exit costs are high; companies tend to remain active despite having negative returns;
• rivals are highly committed to the industry. This commitment can, for example, be the result
of historical reasons;
• existing competitors are unable to understand each other’s signals well, because they are
unfamiliar or use different strategies.
Intensity is not the only aspect of rivalry influencing profit margins. Whether or not rivals compete
on a similar dimension has an impact as well. If all rivals compete on price-level, this negatively
impacts on industry wide profit margins. However, if rivals start to compete on different
dimensions, e.g. searching for other markets trying to serve the needs of different customer
segments, rivalry can have a positive effect on profit margins.
3.3.6 Sixth force
In reaction to the establishment of Porter’s five forces, several authors tried to integrate additional
forces. For example, Grove (1996) proposed the government as a sixth force. Porter (2008),
however, argues that the influence of the government is already integrated in the five forces. By
exploring the five forces, government policies can be analyzed.
3.3.7. Relationship between IIDs’ features and Porter’s forces
The model of Porter’s five forces explains differences in company performance. The explaining
features of IIDs are integrated in the five forces; Table 3.4 presents the different features in relation
to Porter’s five forces.
Becattini’s distinctive features
Porter’s five forces
- labor market
- suppliers of labor
- industrial organization
- suppliers of materials and buyers of goods
- industrial organization
- rivalry amongst existing companies
- institutions
- new entrants
- financial resources
- new entrants
Table 3.4 Becattini’s distinctive IIDs features corresponding to Porter’s five forces
Source: Becattini (1990) and Porter (2008)
3.4 Company performance measurement
Company performance of industrial districts is the dependent variable of present research. Its
measurement and related issues are addressed throughout this paragraph.
Company performance can be measured using several indicators. Each indicator can exclude
24
distorting influences on the data. Manipulation by managers, company size, and non-registered
labor are examples of such influences. This might create a distorted view on company performance
and, consequently, may result in bias. For this reason, existing studies on company performance
usually consider several indicators simultaneously. Throughout the next paragraphs, different
company performance indicators are introduced and discussed. Thereafter, related literature using
company performance indicators for IDCs is presented.
3.4.1 General research on company performance
Brunello et al (1999) make use of operating income growth and EBITDA to measure the
performance of Italian listed companies. EBITDA is defined as earnings before taxes, interest,
depreciation, and amortization. This indicator cannot be easily manipulated by managers, since it
does not incorporate the choice of the depreciation and amortization regime and tax regime.
However, this variable does not account for company size.
One way of taking into account company size is integrating the number of employees in the
measurement. Company productivity can be expressed by profit per employee (PPE). This measure
divides earnings before tax and interest (EBIT) by the total number of employees. This is presented
in Equation 1.
(1)
PPE reveals company efficiency controlling for company size. This allows comparing PPE among
companies of different size. This profitability measure focuses on the human capital of a company,
namely its employees. However, the question remains if PPE is the result of lower wages or higher
employee productivity.
Brealey et al (2004) describe return on assets (ROA) as a commonly used performance indicator,
which is defined as the ratio of net income over total average assets.8 This measurement
successfully combines the profit generating strength of the company’s resources (profit margin)
with the degree of efficient use of the company’s assets (asset turnover) (Gray & Needles, 1999)
and presented in Equation 2.
(2)
Earnings resulting from goods sold after deduction of interest and tax payments are defined as net
income. In Equation 2, paid interest is included in net income, because profits are divided between
debtholders and shareholders and interest payments to debtholders do not imply the company
being less profitable. As a fraction, the ROA performance indicator controls for company size. A
higher ROA refers to companies being more effective in generating income from their assets in
place. When analysing a companies’ ROA, it is important acknowledging that companies can
improve overall profitability by increasing profit margin, asset turnover, or both.
8
Average of assets at beginning and end of time period are used, since assets’ value are likely to change over time
25
Return on equity (ROE) is related to shareholders’ profit. It is measured by the ratio of net profit
over average shareholders’ equity. This is presented in Equation 3.
(3)
!!
" #
ROE reveals how much profit a company generates with the funds invested by shareholders. ROE
can be used for comparing companies’ profitability within the same industry.
3.4.2 Research on company performance in IIDs
One of the scholars measuring industrial district performance using company-level performance
indicators is Luigi Federico Signorini (1994). Using balance sheets of 500 small and medium sized
industrial companies active in the wool sector, Signorini compares company performance within
two different industrial districts with isolated companies. The study uses return on investments
(ROI), value added per employee, and ROE as performance indicators. ROI and ROE are rather
financial profitability indicators, while value added per employee is a productivity indicator. The
research concludes that IDCs perform better on the indicators ROI and value added per employee.
ROE results higher for IDCs as well, however, less consistently than the other indicators. This could
be the result of different financial circumstances of the companies in the districts.
Signorini argues that companies have a strong incentive for ‘reverse window dressing’
(underreporting profits in order to avoid or evade corporate taxation), which might distort the use
of financial accounting data. In the context of this research, however, the only relevant distortion
would be one that would have differential effects depending on company location (companies
either inside or outside IIDs). So far, there are no scientific indications that such differentials exist.
Next to profit underreporting, Signorini mentions the unreliability of capital stock values (including
asset valuation) in accounting data, which should be considered if used for empirical purposes.
Nevertheless, just as with profit underreporting, the same conjecture applies regarding company
location related differences to this regard: as long as differences apply to all companies in the
sample, those distortions are irrelevant for observing differences between IDCs and non-IDCs.
Fabiani et al (2000) expand this initial research, analyzing ROI and ROE in 13 economic sectors
using panel data from 1992-1995. They find significantly higher ROI and ROE for IDCs in nearly all
sectors under study. The observed difference between IDCs and non-IDCs is specifically substantial
for small sized companies.
To summarize, numerous possible company performance indicators exist. Each of them excludes
different potential biases. Table 3.5 presents performance indicators and their characteristics.
Performance indicator
Control
Size
Managerial
Focus
Earnings
Efficiency
EBITDA
√
√
PPE
√
√
ROA
√
√
√
ROE
√
√
Table 3.5 Company performance indicators and their characteristics
26
Most of them are applied in earlier studies on company profitability of Italian industrial districts.
This study considers multiple performance indicators to assure providing a comprehensive image
on company performance. The use of multiple performance in
indicators
dicators rules out the possibility of
one single indicator distorting the results and, consequently, the research’s conclusions.
Because of data availability and the specific factors under study, this research uses (1) EBITDA, (2)
PPE, and (3) ROA as measures
ures for company performance.
3.5 Hypotheses and propositions
This paragraph presents a number of propositions and hypotheses based on the inferences from
the theoretical literature discussed in previous sections. Hypotheses are empirically tested using
company level financial accounting data. Propositions are evaluated along information obtained
during in-depth
depth interviews with different actors from the Fermano
Fermano-Maceratese
Maceratese footwear district.
Related features from paragraph 3.2 are visualized in Figure 3.2.
Figure 3.2 Features related to company performance
3.5.1 Labor market
The availability of a specialized pool of employees in industrial districts is already mentioned by
Marshall and presented as one of the reasons for companies to agglomerate (Marshall, 1890). The
presence of skilled workers in the proximity of companies re
reduces
duces time and effort required to
search, find, select, and finally employ a new worker. This results in lower transaction costs for
IDCs with respect to recruitment and selection.
However, Marshall acknowledges
owledges possible pitfalls of the presence of a highly
ly specialized labor pool
as well. A geographically bounded labor demand may induce companies to compete for skilled
workers by offering specialized workers increasingly higher wages. Paniccia (2002) identifies a
shortage of qualified personnel in all 39 d
district
istrict explored. The shortage of qualified personnel leads
to ‘labor poaching’,’, in which companies offer higher wages to take away highly qualified employers
from competing companies. The particular feature of industrial districts regarding the presence of
o a
skilled labor pool, thus may have a downside as well.
Prospect
rospect workers are aware of the fact that possessing specific skills is remunerated and invest in
training to acquire such skills consequently (Casavola et al, 2000). In the end, the equilibrium wage
27
level in IIDs is determined by supply and demand for labor and due to the factors aforementioned
this may result in higher wages and average labor costs for IDCs accordingly. However, as long as
wage increases -wage being the marginal product of labor in economic theory- are offset by a
proportional increase in labor productivity, higher labor costs cancel out on a net base.
Next to the theorem of higher average labor costs in IIDs, a contrary theory of lower labor costs in
IIDs exists. This latter argues that due to the presence of unregistered labor, companies pay fewer
employer taxes, which reduces total average labor costs. Hadjimichalis (2006) mentions that such
occurs especially among small companies in IIDs. According to Murat & Paba (2004), lower labor
costs in IIDs are especially created by low-skilled labor supply by immigrants from developing
countries. With regard to this discussion, the question arises to what extent this would be
exclusively the case for IDCs rather than a common practice throughout entire Italy, however.
Brusco (1990) observes increasing wage levels in IIDs from the 1950’s until the 1980’s (see Table
3.3): where average wages are below national industry averages from the 1950s until the mid
1970’, IID wages are on average industry wide levels from the 1980’s onwards. In their research, De
Blasio & Di Addario (2005), address the question whether employees in IIDs on average obtain
higher wages compared to employees outside IDs. They extract data from the survey of household
income and wealth from the Banca d’Italia (central bank of Italy) investigating wages of people
active in 233 sistemi locali di lavoro (local labor market areas or LLMAs), of which 28% IIDs. The
results indicate that workers inside industrial districts have a higher chance of finding employment.
However, there appears to be no difference in average wage between workers inside and outside
IIDs.
This is not in line with research by Solinas (1982) and Casavola et al (2000), who use 1989 INPS9
data showing that IID workers’ investment in human capital is rewarded by higher wages
compared to non-IID workers.
Having discussed recruitment related transaction costs and labor costs, the following proposition
and hypothesis regarding labor in industrial district companies (IDCs) are presented:
Proposition (1)
IDCs obtain faster suitable employees than non-IDCs
Hypothesis (1a)
Hypothesis (1b)
Labor costs are higher for IDCs compared to non-IDCs
Labor costs are positively related to IDC performance
3.5.2 Industrial organization
Coase (1937) discusses the organization and the boundaries of the firm. The organization of the
firm relates to (1) the internal structure of companies and its relation to the market structure and
(2) the interaction (horizontal and vertical to the production chain) of companies in forms of interfirm competition and cooperation. Becattini regards competition and cooperation as a specific
feature of industrial districts. The boundaries of the firm relate to the choice companies face when
deciding which activities to carry out (1) internal or (2) external to the company.
Coase’s work on the boundaries of the firm is elaborated by 2009 Nobel Prize laureate Oliver
Williamson’s transaction costs theory (1975). Transaction costs theory adopts a contractual
approach to economic organization and is based on the characteristics of transactions. It argues
9
Istituto Nazionale Previdenza Sociale, Italian national institute for social insurance
28
that a company should integrate those activities if transactions are characterized by (1) a high
degree of asset specificity,
y, (2) a high degree of uncertainty, and (3) of high frequency on the one
hand and that a company should use the market for transactions characterized by low levels of
asset specificity, uncertainty, and frequency on the other hand. Dei Ottati (1987) integrates
integ
transaction costs theory to explain how forces of competition and cooperation work within IIDs
using the concept of ‘normal
normal cooperation
cooperation’,’, which is described in detail throughout the following
paragraph. Figure 3.3 graphically synthesizes transaction costs theory showing the optimal
governance mode for increasing levels of asset specificity.
Figure 3.3 Transaction governance
Based on: Williamson (2002) and Dei Ottati (1987)
Figure 3.3 demonstrates that for low levels of asset specificity market tra
transactions
nsactions are optimal,
whereas for high levels of asset specificity vertically integrating the transaction is optimal. For
intermediate levels of asset specificity, finally, normal cooperation is proposed as optimal
governance mode.
3.5.2.1 Cooperation and
d competition
The interplay between cooperation and competition is assessed by Dei Ottati (1987, 1994).
Industrial districts are characterized by a large number of SMEs specialized in a certain phase of the
industry’s production process. In order to assure industry performance in IIDs, coordinating its
independent companies is necessary. This coordination takes place through cooperation. The
degree of required cooperation between companies in IIDs depends on the type of transactions.
Transactions in which opportunistic
pportunistic behavior would not lead to potential disproportional large
losses can be coordinated with, in the words of Dei Ottati, normal cooperation, which reflects
cooperation based on local customs. Local business customs are known and respected by local
loc
agents in IIDs, because violation can lead to social sanctions and loss of reputation. Therefore, the
use of normal cooperation is supposed reducing transaction costs (compare feature 6, paragraph
3.1).
29
Normal cooperation assures the coordination of the majority of transactions within IIDs. Next to
this kind of transactions, transactions in which opportunistic behavior would lead to high losses
exist, for example if dedicated assets are involved. Dedicated assets relate to a form of asset
specificity defined by Williamson (1985) as: ‘investments in production capacity which normally
would not be undertaken, but for the prospect of selling a significant amount of products to a specific
customer’. This kind of transactions requires higher safeguards than normal cooperation can
provide. The neo-classical solution for companies to cope with transactions, in which high levels of
asset specificity are involved, is internalizing such transactions through vertical integration.
Inside IIDs an additional safeguard is provided in the form of personal reputation. Hendrikse
(2003) relates reputation value to the (1) frequency at which transactions occur, (2) the time span
of the inter-company relationship, and (3) the probability of separate transactions. It takes several
transactions for a reputation to be created and established. However, once established, a strong
reputation can generate large future returns. Cooperation is facilitated by reputation and allows
IDCs using market transactions to a greater extent rather than vertical integration. Compared to
IDCs, non-IDCs are forced to rely relatively more on vertical integration. The use of normal
cooperation instead of vertical integration allows for the existence of phase specialization in IIDs,
which increases productivity at company level and the entire industry at district level. However,
cooperation may have a downside as well. If cooperation becomes anti-competitive and companies
collude, cooperation can become destructive for IIDs.
Whereas cooperation is necessary in IIDs to coordinate complementary activities of IDCs,
competition arises in case of substitutable activities of IDCs. Companies located in IIDs do not only
compete with respect to price of their products, but on additional dimensions like specific product
characteristics, organizational structure, and use of new technologies as well. The presence of a
large number of IDCs results in fierce horizontal competition, which forces companies to increase
efficiency and innovative capabilities (Van Dijk, 1995). For example, when a company has
introduced a new machine which improves a company’s productivity, competing companies are
forced to implement this technology as well to prevent them from having higher costs than their
rivals. The high number of substitutable activities increases this mechanism, increasing IIDs’ overall
competitiveness.
Like cooperation, competition can have a downside as well, however. Fierce price competition can
become destructive and reduce profits to such a degree that it drives companies out of the market.
This mechanism is increased in IIDs because of the high number of horizontally competing
companies.
Concluding, through inter-firm cooperation vertical to the production chain transaction costs are
assumed to decrease, which leads to higher performance of IDCs. Inter-firm competition horizontal
to the production chain is supposed forcing companies to be more efficient and enhancing their
innovative capabilities. This leads to higher performance of IDCs as well. Therefore, the following is
proposed:
Proposition (2a)
Proposition (2b)
IDCs exhibit a higher degree of vertical inter-firm cooperation than non-IDCs
IDCs exhibit more intense horizontal inter-firm competition than non-IDCs
30
3.5.2.2 Specialization
The division of labor is regarded one of the distinguishing features of industrial districts (Brusco,
1986; Becattini, 1990). Phase specialization refers to the division of labor such that each company
specializes in one (or few) stages of the production chain and causes the production chain to
vertically disintegrate.10 This enables companies benefiting from economies of scale, since sufficient
aggregate demand for a single stage component from companies within the same industrial district
permits the specialized company to focus on the production of just one single component of the
entire production chain. Such focus commonly leads to higher usage levels of available production
capacity and accumulated experience and, consequently, to increased productivity. Phase
specialization, thus, positively impacts company performance. This research regards specialization
as vertical disintegration, that is, the inverse of vertical integration.
Piore & Sabel (1984) introduce the concept of flexible specialization. As opposed to the Fordist
model of production, which merely relies on the idea of standardized mass production intended for
undifferentiated mass consumption being the leading paradigm in industrial production
throughout the 1940’s-1960’s, flexible specialization allows for customized and differentiated
production in limited batches. Table 3.7 highlights the main differences between flexible
specialization and Fordist production.
Feature
Flexible specialization
Fordist production
manufacturing method
· limited quantities of
differentiated products
· mass production of
standardized products
organization of work
three cooperation levels:
Tayloristic separation:
(1) white-collar workers
(2) skilled workers
(3) unskilled workers
(1) white-collar workers
skill accumulation
· possible for skilled workers
· not possible
production procedures
· customized
· standardized
production organization
· multiple SMEs in IDs
· few, large factories
(2) unskilled workers
Table 3.7 Differences between flexible specialization and Fordist production
Source: Capecchi (1990)
A vertically integrated company creates internally the biggest part of the value added and therefore
exhibits a relatively high relationship between value added and revenue. Contrarily, a (flexibly)
specialized company, carrying out just one or few phases of the production process, is only
responsible for a limited share of the created value added. Therefore, one expects the relationship
between value added and revenue in a specialized company to be relatively low if compared to a
completely vertically integrated company, ceteris paribus (Van Dijk, 1995;Cannari & Signorini,
2000). This is visualized in Figure 3.4: company P on the left is completely vertically integrated and
10
For footwear production generally five different phases are distinguished: (1) cutting leather for the upper, (2)
edging and sewing the upper, (3) montage of upper part on lower part (lower part consisting of sole and heel) and
assembly of other accessories, (4) finishing (refinement and polishing) of shoe, and (5) packaging of final product in
shoeboxes and preparation for distribution (Varaldo, 1988)
31
carries out all production phases internally, whereas the specialized production chain on the right
represents strong phase specialization by independent companies.
Figure 3.4 Degree of specialization determined by value added share
Signorini (2000) uses the ratio between value added and sales to determine the degree of
specialization in the knitwear industry of two Italian industrial districts using 1991 company level
financial accounting data. The results reveal that IDCs are more specialized than non-IDCs.
Conti et al (2007) divide value added over revenue to establish the degree of specialization using
data from Banca delle Marche showing increasing specialization levels among a sample of 79
medium sized footwear companies from the Fermano-Maceratese district in the period 1993-2003.
This present research considers the degree of specialization adopting a similar approach. The
hypothesis expects higher levels of specialization among IDCs compared to non-IDCs.
Hypothesis (2a)
Hypothesis (2b)
IDCs are more specialized compared to non-IDCs
Specialization is positively related to IDC performance
3.5.3 Institutions
Institutions delivering services to companies in Italian industrial districts have received
considerable attention in literature (Rabellotti, 1995). Becattini refers to institutions including both
economic and social institutions. Economic institutions consist of co-operative movements, trade
associations, and labor unions. Social institutions consist of church, school, and family.
Economic institutions (trade associations in particular) represent three important actors in an
industrial district: respectively the workers, subcontractors, and, finally, companies (Dei Ottati,
2002). This study argues that conversation between those actors is necessary at the local level for
three reasons: (1) the development of a long term vision, (2) the protection of labor and suppliers
from exploitation, and (3) the supply of collective goods such as trade fairs and specific training. In
32
addition, Brusco (1990) points out the importance of institutions triggering innovation and
fostering change within IIDs. The presence of institutions and the services they provide are
believed to explain IIDs’ performance (Harrison, 1991). However, research results do not always
confirm the beneficial impact. Brusco & Righi (1997), for example, do not provide ’sound evidence
that creation of institutions has a positive impact on the economic performance of Italian companies‘.
Apart from creating a stimulating environment for the companies, institutions can create an
enabling societal environment as well. An example is the role of a (technical) school. There are
indications that the lack of schools can contribute to the underdevelopment of SMEs in certain
areas (Boari, 2001).
Rabellotti (1995) assesses the impact of institutions by means of interviews with actors of the
Fermano-Maceratese and Brenta-Veronese (northern Italy, see Figure 2.1) footwear districts. The
study indicates that associations of SMEs of the Brenta-Veronese district are well organized. This
resulted, among others, in the creation of a centre of technological assistance and an export
consortium. Despite its larger size, in the Fermano-Maceratese district only one trade fair
organizing centre is founded. Actors in the Fermano-Maceratese district confirm that a lack of
tradition of collaboration and institutional intervention exist in the region. They praise the BrentaVeronese district as best-case practice for the role of institutions benefiting the entire district.
Rabbelotti concludes that local associations of entrepreneurs and local governments are able to
create specialized service centers, which support the industrial sector, but that institutions are not
crucial for districts’ development (Rabbelotti, 1995).
There is no clear consensus regarding the presumed beneficial impact of institutions11 on IDCs.
Despite this, there have been initiatives of local governments to provide services to district
members, consisting of information provision and professional consultancy services (Boari, 2001).
The particular focus on district members would suggest that institutions do not reach those
companies located outside the IID. The propositions are therefore:
Proposition (3a)
Proposition (3b)
Proposition (3c)
Proposition (3d)
Institutions’ services are specifically targeted towards IDCs
IDCs are more often associated to institutions compared to non-IDCs
IDCs make more use of services provided by institutions
IDCs are more satisfied by the services provided by institutions
3.5.5 Financial intermediaries
Generally, companies raise financial resources necessary for investments and daily operations in
two ways: (1) as equity (shareholders and retained earnings) and (2) as debt via financial
intermediaries and/or inter-company lending (commercial lending). The access to external
financing (credit) and the cost and extent of obtaining credit is assumed to be of crucial relevance
for the functioning of companies and, consequently, to their performance.
The Italian reality of financial intermediaries presents a dichotomy between national and local
banks. The presence of local banks is considered a distinctive characteristic of industrial districts
(Becattini 1990; Farabullini & Gobbi, 2000; Alessandrini et al, 2008).
11
Reference is made to economical institutions (especially trade associations); not to social institutions
33
Following Becattini (1990), local banks are likely to provide more credit at relatively favorable
conditions to local companies compared to national banks. This is due to the assumed close
interaction between local banks and the local community, which provides the local bank with up-todate knowledge about the local reality and its entrepreneurs. Under normal circumstances, the
presence of asymmetric information and the possibility of opportunistic behavior by borrowers are
compensated by the bank through a (1) risk premium on top of the interest rate and by (2) credit
rationing (banks limiting the supply of loans to prevent adverse selection). In industrial districts,
however, the local community can support the local bank in evaluating the borrower’s risk before
credit issuance, which reduces possible negative outcomes due to asymmetric information between
borrower and lending bank. The local community assists the bank as well in monitoring the
borrower. This because the community is believed to discourage moral hazard and opportunistic
behavior of individual members, given that such behavior damages the local community as a whole
(Finaldi Russo & Rossi, 2000). Finally, local banks are said to operate more quickly in the credit
issuance process, given the absence of multiple hierarchical layers (as opposed to large national
banks).
An alternative view on the differences between national and local banks argues that local banks
have higher costs in evaluating risks and controlling IDCs, however. According to this vision, local
banks have higher monitoring costs since, in addition to the performance of the single company, the
local bank should assess the performance of the company’s network of subcontractors and other
business partners as well. Another pitfall derives from the local banks’ concentration on a restricted
geographical area, specialized in just one industrial sector. This does not eliminate the diversifiable
credit risk on a wide range of economic activities (Finaldi Russo & Rossi, 2000). Next to that, small
local banks may not enjoy economies of scale to such an extent large national banks do.
Signorini (1994) seems to confirm the suggestions proposed by Becattini (1990). Signorini analyzes
financial accounting data of 500 companies active in the wool sector and concludes that IDCs
receive more credit from banks compared to non-IDCs.
In a similar study, Finaldi Russo & Rossi (2000) use financial data of 1’702 manufacturing
companies between 1989 and 1995 to explore possible differences in leverage (ratio between debt
and equity), interest rate, and the relationship with the banking system between an evenly divided
sample of IDCs and non-IDCs. The results indicate that IDCs have higher leverage ratios, indicating
those companies being more indebted to banks. However, this does not negatively influence
company performance as IDCs demonstrate higher return on assets (ROA).
A final note on leverage: using debt funding is advantageous from a fiscal point of view, since Italian
corporate tax law allows deducting debt interest repayments from taxable income. Theoretically,
this would induce companies to use as much debt as possible to reap the full benefits of this tax
advantage. However, increasing levels of leverage carry along a number of drawbacks as well
(higher agency costs, higher risk of financial distress, and increasing downside of asymmetric
information). Trading of advantages and disadvantages of leverage by a company’s management
would predict a certain optimum level of leverage (Modigliani & Miller, 1958; Grinblatt & Titman,
34
2002; Brealey et al, 2004).
So far, the role of financial intermediaries in industrial districts in general and the functioning of
local banks in particular are reviewed. Leverage has been addressed and is assumed to
demonstrate higher levels among IDCs compared to non-IDCs. This leads to
Hypothesis (3a)
Hypothesis (3b)
IDCs have higher leverage ratios compared to non-IDCs
Leverage is positively related to IDC performance
Proposition (4)
IDCs make more often use of local banks compared to non-IDCs
35
4. Methodology
In order to facilitate an analysis on company performance of the Italian footwear sector, it is
necessary establishing an appropriate methodological approach. This chapter describes the
methodological approach deployed in this thesis.
The research question addresses two sub questions. The first question concerns a comparison of
company performance between IDCs and non-IDCs. An answer is provided using a quantitative
analysis. The second sub question asks why the performance indicators of IDCs and non-IDCs
eventually differ and is assessed according to the hypotheses and propositions as developed
throughout the preceding chapter.
In order to acquire insights into these complex relationships a holistic view is necessary. The use of
a case-study can provide this view (Baarda et al, 2005). Case studies are defined by Yin (1994) as:
’An empirical inquiry that investigates a contemporary phenomenon within its real-life context, in
which the boundaries between the phenomenon and context are not clearly defined. The inquiry copes
with the technically distinctive situation in which there will be many more variables of interest than
data points, and as one result relies on multiple sources of evidence, with data needing to converge in a
triangulating fashion, and as another result benefits from the prior development of theoretical
propositions to guide data collection and analysis.’
Case studies are able to present a comprehensive explanation of a situation. It is a preferred
strategy for studies focusing on the ‘how’ and ‘why’ questions surrounding a certain topic. Case
studies make use of different types of data sources. The data collected in this research consist of
both quantitative and qualitative data. The quantitative data in this study are acquired from a
database. The qualitative data are acquired by means of interviews. This chapter is divided in two
parts: the first describes the quantitative method, whereas the second provides a description of the
methodological analysis of the qualitative data.
4.1 Quantitative method
This research uses quantitative methods to distinguish possible differences in company
performance between IDCs and non-IDCs at first, as well as for investigating possible explanations
for any possibly observed differences at second. First, the origin and selection of the used data are
discussed. Hereafter the generation of dependent, explanatory, and control variables are presented.
Finally, the applied statistical methods are described.
4.1.1 Dataset description
The data are obtained from the Amadeus database. This online database contains both financial
information and business intelligence about 13 million European companies. Since company
financials are inserted in a standard format, comparison between companies is highly feasible.
Amadeus allows its users deploying various queries based on a range of search criteria and filters.
In addition to geographical criteria, it enables identifying companies by means of industry
36
classification codes. The international classification standard is called NACE and has a maximum
number of four digits. For the purpose of this research, NACE Rev. 2 1520 (the classification for
footwear manufacturing) is used. This classification includes producers of intermediate goods like
soles, heels, shoelaces etc. s well. It excludes auxiliary activities (for example: machinery and
equipment for footwear manufacturing, packaging, tanneries). Using both geographical and
sectorial filters, an initial dataset consisting of 3’511 observations is extracted.
4.1.2 Variables description
This paragraph presents the variables used for the empirical testing of the hypotheses. It
distinguishes for (1) dependent variables, (2) explanatory variables, and (3) control variables.
4.1.2.1 Dependent variables
The dependent variable of the analysis is company performance and is assessed using performance
indicators described throughout paragraph 3.4: (1) EBITDA margin, (2) return on assets (ROA), and
(3) profit per employee (PPE). The use of these indicators provides a complete view, since they
exclude different potential biases and focus points (compare Table 3.1). The indicators are directly
obtained from the data provided by Amadeus.
4.1.2.2 Explanatory variables
In order to be able to attribute differences in company performance based on location, it is
necessary to divide the entire group of observations into two subgroups based on their respective
location either inside or outside an area formally defined as industrial district. Therefore, a
dichotomous dummy variable is created, indicating if a company is located inside or outside an IID.
Defining IIDs is done by applying the so called Sforzi-ISTAT classification.
Sforzi (1989; 1990) identifies IIDs in a two step process. The first step consists of the identification
of so called local labor market areas (LLMAs or sistemi locali di lavoro). These are areas in which
large shares of employees commute from their residence to the place they work.
The second step consists of determining whether in an LLMA (1) the share of employment in
manufacturing industries over total employment is higher than the Italian national average, (2) the
share of employment in manufacturing industries in companies with less than 250 employees is
higher than the Italian national average, and (3) the share of employment in one or several
manufacturing industries is higher than the Italian national average. If all conditions are met, an
LLMA is considered as an IID (leaving the possibility of single IIDs consisting of several LLMAs).
Although the definition of IIDs by Sforzi is the one most commonly used in IID literature, it is not
comprehensive and carries along a number of pitfalls, which are discussed in detail in chapter 6.
ISTAT (2005) uses Sforzi’s method to define IIDs based on 2001 census data. The ISTAT definition
geographically delimits industrial districts based on postal code and municipality. Combining
information on postal code and municipality, this research applies the same methodology in
determining if a company belongs to an area defined as industrial district.
Next to the variable district, which is actually a grouping variable dividing the entire dataset in two
subgroups of IDCs and non-IDCs, three further explanatory variables are considered by this study.
Labor costs are directly obtained from the information offered by Amadeus. They refer to the
average annual labor costs per employee expressed in thousands of Euros. Companies may
37
underreport the true number of employees for fiscal reasons, as indicated in paragraph 3.4.2 on
reverse window dressing. With regard to differences between IDCs and non-IDCs the same holds
here as well: as long as both IDCs and non-IDCs underreport the number of employees, the possible
distortion is symmetrical to both subgroups and does not influence the results.
Specialization is approximated by taking the inverse of vertical integration. For measuring vertical
integration, two conditions must hold: (1) ‘it must be an extension of, and consistent with, accepted
economic doctrine’ and (2) ‘it must be operational and capable of statistical measurement’ (Adelman,
1955).
Generally two different strategies exist: (1) a ratio measuring value added over revenue or (2) a
vertical ratio approximating the percentage of total product which is part of a company’s vertical
chain. The first approach is commonly used in empirical research, given the advantage of ease of
calculation and data availability. However, the first approach presents two serious shortcoming: the
ratio is easily affected by factors other than actual vertical integration (profitability in particular)
and ‘the ratio is greater the nearer the company is to an extractive process, because value added is
usually relatively greater at a primary level’ (Maddigan, 1981).
For the second approach, again two drawbacks are identified. The first drawback is of theoretical
nature, as the ratio reflects both vertical and horizontal integration. This means that the ratio can
be distorted upwards as the result of horizontal mergers. The second drawback relates to practical
applicability, since data required for constructing such a ratio are usually difficultly available, as
they require a breakdown of the total value of production by production line (Maddigan, 1981).
Because of appropriate data availability, this research approximates specialization using the
inverse of the ratio of ‘value added12’ over ‘revenue13’. Both items are directly provided by Amadeus
from the annual profit and loss statement and expressed in thousands of Euros. Applying this
operationalization for specialization, is identical to the one applied in prior IID research (Signorini,
2000; Conti et al, 2007).
Leverage is commonly estimated by dividing long term debt over equity (Wippern, 1966; Dhaliwal,
1986). There are two possible applications of this ratio: (1) at book values or (2) at market values,
both of which contain important conceptual biases. The ratio of debt to equity at book value
measures the relationship between the par value of debt, plus or minus any unamortized discounts
or premiums, to the amount of equity as determined by the historical costs of assets less the book
value of outstanding debt. This ratio, then, reflects the accumulation of all past accounting valuation
and income determination decisions. Its use as a measure of financial risk implies that the
relationship between debt and the book value of equity is the one which is most relevant in
determining lenders' and investors' claims to the earnings stream of the company. This implication
is justified only if one is willing to accept the view that the risks of fixed charge financing arise
solely out of the potential losses incurred in liquidation and that book values adequately reflect that
loss potential (Wippern, 1966).
The measure of leverage based on market values overcomes some of the difficulties discussed above.
12
Value added is defined as the sum of all taxes paid by the company during the accounting period plus profit after
taxation and all employees costs of the company (including pension costs) plus extraordinary and other profit plus
total amount of depreciation and amortization of the assets plus total amount of interest charges paid for shares or
loans
13
Revenue is defined as the sum of net sales, other operating revenues, and stock variations. The figures do not
include value added taxes, excises taxes, or similar obligatory payments
38
There are, however, other problems associated with the market value ratio. Modigliani & Miller
(1958) refer to the problem of bias resulting from the division of both the dependent and
independent variables by the market value of equity in the correlations they report. This manner of
using market value does introduce a statistical bias into the correlation. An even more important
conceptual bias is introduced by measuring leverage as the ratio of debt to the market value of
equity. It is generally recognized that the market value of the equity of a company is a function of a
number of variables in addition to financial structure. If a company's shares are given a relatively
high value by the market for reasons other than, or in addition to, its financial structure, that
company's leverage ratio will, of course, be lower than it would if its shares were less favored by
investors (Wippern, 1966).
For approximating leverage, this research takes the ratio of ‘noncurrent liabilities14’ over
‘shareholders funds15’. Both are balance sheet items provided in Amadeus and expressed in market
values (thousands of Euros). Using market values is not opted for specifically, but just applied
given data availability from Amadeus.
4.1.2.3 Control variables
Next to the explanatory variables described in the preceding paragraph, company performance can
be influenced by several other factors as well. Those factors should be taken into consideration and
controlled for when applying quantitative methods and empirical testing. That is insured for by
including of control variables (Gujarati, 2003).
This research uses the following four controls (1) company age, (2) company size, (3) location and
(4) legal status. Company age is expected to be positively related to company performance, as
experience accumulated over time enhances companies’ efficiency. However, this supposed
positive relationship may be moderated by a phenomenon called ‘leapfrogging’, a notion introduced
by Joseph Schumpeter (1942) referring to incumbent companies having less incentive for
innovation. With respect to company age, older companies being in a disadvantageous situation
may be outperformed by newcomers from a second generation (Cucculelli & Micucci, 2008). The
variable company age is constructed using companies’ date of incorporation and expressed in
integer number of years. Accounting for company size is trivial in econometric research on IIDs
(Signorini 1994; Fabiani et al, 2000; Iuzzolino, 2000). Company size is a dichotomous dummy
variable distinguishing between SMEs and larger companies. For defining the size of SMEs, this
research adopts the approach provided by Amadeus and considers companies with an annual
turnover lower than €15 million as SMEs.
As indicated in paragraph 2.1, in Italy a significant difference in the economic structure between
southern regions and the remaining regions exists (that is, southern regions are economically less
developed). In order to take into consideration this particularity of the Italian economic structure,
a control dichotomous dummy variable southern Italy is created, indicating if a company is located
within one of Italy’s southern regions16.
Legal status is the final control variable used in the empirical part and takes into consideration the
14
In Amadeus, noncurrent liabilities are defined as the sum of long term financial debts to credit institutions (loans
and credits) plus all long term liabilities not related to financial institutions (taxes, group companies , pension loans)
15
Shareholders funds are defined as the sum of issued share capital (authorized capital) plus all shareholders funds
not linked to issued capital (reserve capital, undistributed profit, and any minority interests)
16
Abruzzo, Basilicata, Calabria, Campania, Molise, Puglia, Sardegna, and Sicilia
39
possibility of a company finding itself in liquidation or bankruptcy procedures. Excluding such
companies would distort the dataset towards solely more profitable and successful companies.
Table 4.1 below summarizes the different variables presented throughout the previous paragraphs.
Variable
Description
Dependent
Y1
EBITDA margin
Y2
return on assets
Y3
profit per employee
Explanatory
X1
district
X2
labor costs
X3
specialization
X4
leverage
Control
X5
age
X6
size
X7
southern Italy
X8
legal status
Table 4.1 Variables overview
Measurement
Operationalization
ratio
ratio
monetary1
(EBITDA) / (revenue)
(net income) / (average total assets)
(net profit) / (number employees)
dichotomous
monetary1
ratio
ratio
0= non-IDC; 1= IDC
(total labor costs) / (number employees)
1 – (value added / revenue)
(long term debt) / (equity)
interval, integers
dichotomous
dichotomous
dichotomous
company age in years in 2007
0= no SME; 1= SME (revenue ≥ €15 mln)
0= other location; 1= southern Italy
0= in liquidation or bankruptcy; 1= active
1: monetary units: x €1’000
4.1.3 Dataset selection
Before empirical tests are executed, one must assure that factors possibly distorting the results are
excluded. Examples are outliers and unrealistic values, which could create bias. Therefore, the
dataset is modified, which is visualized in Figure 4.1
Figure 4.1 Dataset selection
The initial dataset contains information regarding 3’511 companies. In order to exclude the
influence of time, only data from the same reference period (calendar year 2007) are used. By using
exclusively 2007 data, possible distorting
istorting influences caused by the 2008 global economic crisis are
40
not taken into consideration. 589 observations do not provide financial accounting data for 2007
and are excluded. The remaining 2’922 observations are assessed with respect to missing values.
This implies that for all observations is verified that all variables (Table 4.1) provide realistic values.
If one or more values are missing, the entire observation is removed; this is the case for
1’717obervations. Finally, the remaining observations are scanned for outliers. Outliers are
detected numerically using skewness and kurtosis indicators and visually using boxplots. Next to
this, values greater than two times the standard deviation are excluded. This leads to the removal of
40 observations.
Selecting data according to the procedure described in this paragraph results in a final dataset
comprising 1’165 companies, of which 538 (46%) represent IDCs and 627 (54%) non-IDCs.
4.1.4 Empirical methodology
The following sections discuss the statistical methods deployed for assessing the hypotheses.
4.1.4.1 Independent sample t-tests
An independent samples t-test provides empirical evidence for potential differences of means
between two groups. It analyzes the differences between the overall means of groups and compares
them to the expected differences between groups. The standard deviations of the groups provide
insights in the expected differences (Field, 2005). If standard deviations are high, large differences
between the groups’ means can occur without indicating a statistically significant difference. If
standard deviations are low, large differences between groups indicate a significant difference.
This research applies independent sample t-tests to establish potential statistically significant
differences in the dependent variable, company performance. For this purpose, the district dummy
is used as grouping variable, dividing the entire dataset into two subgroups (IDCs and non-IDCs
respectively). Next to the dependent variable, t-tests are used to determine possible differences
between IDCs and non-IDCs for the explanatory variables as well.
4.1.4.2 Ordinary least square regression
A linear regression can provide empirical evidence for causal relationships between dependent and
independent variables (Wooldridge, 2003). This study uses ordinary least square estimation (OLS)
to test the hypotheses presented in paragraph 3.5. The assumptions for OLS estimation are
presented in Table 4.2. The tests indicated in the third column are applied for the empirical testing.
Assumption
Linearity
Absence of
multicollinearity
Homoskedasticity
Description
· relationship between Yi and Xi linear
· no exact linear relationship between
two or more independent variables
· error term ε has constant variance
for all observations:
E$% =& %
Absence of
· random variables $ statistically independent:
E($ $' )=0 for all i≠j
serial correlation
Normality
· error term $ normally distributed
Table 4.2 Assumptions ordinary least squares estimation
Source: Pindyck & Rubinfield (2003)
Test
·Variance Inflation Factor
· Breusch - Pagan
· Levene
· White
· Durbin-Watson
· Jarque-Bera
41
Equation 4 represents the regression model applied in present research. It investigates the impact
of the explanatory variables (as spelled out in Table 4.1) on company performance, while
controlling for age, size, location, and legal status.
( )* + )% ,% + )- ,- +. . . +)/ ,/ + $
(4)
The model of Equation 4 represents solely linear relationships. In order to integrate possible
increasing or decreasing marginal effects of the explanatory variables, quadratic functions of the
explanatory variables may be included. Significant quadratic functions with coefficients of opposite
signs (compared to the original model without quadratic terms) imply a diminishing effect of
explanatory variable on dependent variable (Woolridge, 2003).
Another point of interest forms the possibility for an explanatory value to demonstrate a
differential effect caused by another explanatory variable. Therefore, interaction terms are
generated by multiplying two explanatory variables. If interaction terms prove to have a significant
effect on the dependent variable, they are included in the model (Wooldrige, 2003).
The creation of interaction terms of explanatory variables and district dummy allows assessing the
net effect on company performance, while distinguishing between IDCs and non-IDCs for the
magnitude of such effect. The empirical results are obtained using the statistics software package
SPSS, version 15.0.
4.2 Qualitative method
The following sections explain the applied methodology for obtaining and elaborating the
qualitative data used in present research.
4.2.1 Data description
The qualitative data are collected through semi-structured interviews conducted throughout June
2008. A total number of 29 interviews is conducted. One should notice that the financial accounting
data used in the empirical analysis refer to a different reference period (2007). This might create
discrepancy between the empirical analysis and the insights obtained during the interviews.
4.2.1.1 Corporate interviews
A first series consists of 22 corporate interviews (CI) with (founding) owners or managers of
footwear companies located both inside and outside footwear districts in the Marche and Umbria
regions of central Italy.
Possible corporate interviewees are selected from Amadeus. Selection criteria refer to the desire
that companies to be approached for interviews reflect the general company characteristics as
those from the Amadeus sample. In addition, for practical matters, the location of the company is
considered as it should not be too distant from the authors’ provisional location in the Marche
region during their stay in June 2009. The corporate interviewees are predominantly approached
by telephone first, after which an appointment for an interview is arranged, in most of the cases
during the subsequent week.
Another part of the corporate interviews is arranged on the spot: the researchers present
42
themselves directly at the location of the company asking for participation after having explained
the purpose of the research, the characteristics of the questionnaire, and after assuring
confidentiality. The response rate is high, as only two out of thirteen approached companies (both
by telephone and on the spot) expressed the wish not to participate in the research.
Next to the companies selected from Amadeus, the researchers are introduced through two local
contacts for a number of interview appointments (seven and four respectively). It should be noticed
that this may cause bias in the selection of the interviewees. The groups of companies contacted via
recommendation by local contacts may form a (possibly non-representative) sub-network within
the entire population of footwear manufacturers in the Fermano-Maceratese area.
Usually one asks for an interview with the founder, manager, or director, since those are expected
possessing the most relevant knowledge related to the topics and are therefore considered the
appropriate persons in answering the questionnaire. This coincides with the authors’ experience:
the majority of interviews (16 out of 22) are conducted with the founder or a direct relative of the
founder and those persons possess the knowledge required to answer the interview questions
adequately.
All corporate interviews are conducted in a circa one hour face-to-face contact setting, which allows
the researcher asking for comprehensive questions (Baarda et al, 2005) and held at the location of
the interviewee. The interviews are conducted in Italian (20) or in English (2) depending on the
preference and ability of the interviewee. The interviews are recorded with a voice recorder for
further processing after consent of the interviewee. One should notice that not using ones mother
tongue (both by interviewer and interviewee) might hinder communication in general and limit the
transmittance of information during the interviews.
The corporate interview questionnaire (see Appendix 9.1) is structured along a number of topics of
interest to the researchers. In order to create a suitable atmosphere in which the interviewee feels
comfortable to freely answering the questionnaire, the interview starts with a series of general
introducing questions. These regard the company’s main economic activity, its main products, the
role of the interviewee within the company, and the foundation of the company. After the
introducing questions, each further topic is introduced by a general question after which more
focused and eventually sensitive questions (for example: competition, financial structure,
relationship with financial intermediaries) are posed. Comparing information available in Amadeus
with the answers of the interviewees allows, to a certain degree, for controlling the correctness of
the provided answers. All possibly verifiable obtained information proves to be correct. In order to
attempt grasping the most salient issue of this research, the questionnaire ends with the location
specific question: ‘Compared to a similar company in (resp. outside) an agglomeration of footwear
producers, which specific advantages and disadvantages do you experience?’
Table 4.4 presents an overview of the corporate interviewees and some basic information of the
interviewed companies.
43
Corporate interviews Fermano-Maceratese footwear district
Interviewee
1)
2)
3)
4)
5)
6)
7)
8)
9)
10)
11)
12)
Employees1
founder / manager
founder / manager
founder / manager
sales manager
founder / manager
founder / manager
founder / manager
business developer
managing director
founder / manager
founder / manager
founder / manager
90
25
30
25
50
4
65
25
160
36
25
4
Annual
turnover2
15.8
unknown
unknown
7.0
unknown
unknown
11.7
unknown
77.1
7.9
unknown
0.6
Main product
male footwear
female footwear
female footwear
female and male footwear
female and male footwear
horseriding boots
female footwear
children footwear
female and male footwear
leather soles
heels for female footwear
uppers for female footwear
Foundation
1969
1990
2003
1980
1976
2003
1977
1985
1971
1979
unknown
1995
Municipality
Province
Monte San Pietrangeli
Montecosaro
Sant’Elpidio a Mare
Torre San Patrizio
Monte San Giusto
Casette d’Ete
Fermo
Monte Urano
Civitanova Marche
Loro Piceno
Monte Urano
Montecosaro
Fermo
Macerata
Fermo
Fermo
Macerata
Fermo
Fermo
Fermo
Macerata
Macerata
Fermo
Macerata
Corporate interviews Ancona and Perugia provinces
13)
14)
15)
16)
17)
18)
19)
20)
21)
22)
financial manager
daughter founder
founder / manager
modeler /
grandson founder
purchaser
son founder
brand manager
managing director
founding partner
business developer
daughter founder
administrator
founder / manager
Table 4.4 Corporate interviews
94
39.9
16
15
0.5
unknown
6
2.8
125
26
15
20
30.8
1.2
unknown
1.9
10
10
0.9
0.2
female footwear
1981
Serra de’ Conti
Ancona
sandal uppers
indoor slippers, sandals,
thongs
synthetic soles
2005
1999
Osteria
Fabriano
Ancona
Ancona
1992
Gualdo Tadino
Perugia
female footwear, apparel
female footwear assembly
female sandals
female footwear uppers,
accessories
shoelaces
female footwear uppers
2001
1987
1993
1994
Piticchio
Serra de’ Conti
Nocera Umbra
Serra de’ Conti
Ancona
Ancona
Perugia
Ancona
1981
2005
Serra de’ Conti
Castelleone di Suasa
Ancona
Ancona
1: full time equivalents (June 2008)
2: in € million (2007)
44
Table 4.4 demonstrates a number of key data for each interviewed company. In the majority of the
cases, the founder /manager is interviewed. The number of employees per company appears higher
in the Fermano-Maceratese district. Annual turnover is indicated if available from Amadeus. A
number of companies interviewed are not available in Amadeus. Companies from the FermanoMaceratese district are considerably older than the other companies.
4.2.1.2 Expert interviews
Besides the corporate interviews, a series of expert interviews (EI) is conducted with
representatives from institutions, financial intermediaries, and professors from a local university.
These actors are expected providing a complete and in-depth view with respect to their specific
field of expertise. Their information and inputs complement the information obtained during the
corporate interviews.
The expert interviewees are selected after an internet research and contacted via email and/or
telephone (5) as well as on the spot (2). The expert interviews are held both in Italian (4) and
English (3). Table 4.3 presents the expert interviewees and their respective expertise fields.
Interviewee
Affiliation
Municipality Province
1)
2)
professor applied economics
researcher, statistician
Ancona
Ancona
Ancona
Ancona
3)
company representative
banker
researcher, sociologist
Monte San
Giusto
Ostra Vetere
Ancona
Macerata
4)
5)
Ancona
Ancona
6)
policy maker, responsible for
innovation and competitiveness
professor industrial organization
university of Ancona
research centre trade association
for manufacturing companies
trade association
for manufacturing companies
local cooperative bank
research centre confederation for
artisanship and SMEs
regional government
Ancona
Ancona
Ancona
Ancona
7)
university of Ancona
Table 4.3 Expert interviews
4.2.2 Data processing
The interviews contain different topics related primarily to the propositions presented throughout
paragraph 3.5. Each topic deals with a feature explaining possible differences in company
performance. The propositions regarding (1) labor market, (2) industrial organization, (3)
institutions, and (4) financial intermediaries are operationalized, which forms the starting point for
the formulation of the questions in the questionnaire. Table 4.5 presents the operationalization of
the propositions from paragraph 3.5.
45
Concept
Labor market
Functioning of labor market
Dimension
Time
availability of suitable labor force
Industrial organization
Vertical inter-firm cooperation
nature of transactions between
companies active in consecutive
phases of production chain
Indicator
· average time required for employing
new personnel
· average duration buyer-seller relationship
Transaction
governance
· sensitivity for disturbances
· degree of knowledge exchange
· use of penalties in case of contract breach
· number of known competitors
Horizontal inter-firm competition
Intensity
· perceived relationship with competitors
beneficial impact trade associations
Range
Impact
Perception
· geographical location of institutions’ members
· number of services used by companies
· perceived impact of institutions
Financial intermediaries
Use of local bank
Access
· category of bank used by companies
nature of relationship between
companies active in similar phases
of production chain
Institutions
Role of institutions
access to services provided by
financial intermediaries deeply
rooted in the local territory
Table 4.5 Operationalization of propositions
This table forms a framework for the questionnaire for the companies, which is attached in
Appendix 9.1. The questionnaire is customized for those expert interviews with actors from
institutions, financial intermediaries, and universities. For example, an interview with a
representative from a trade organization will be primarily focused on the propositions regarding
trade organizations.
Because of the use of interviews, the qualitative data can be influenced by subjective actions. This
effect may be enhanced by the open character of the interviews (Baarda et al, 2005). The recorded
interviews are translated where necessary and typed out. Adopting this approach rules out the
possible threat for bias presented by subjective memory of the researcher. The final datasets are
created after scanning the raw data and deleting non-relevant information (for example:
interruptions by phone calls or information related to the interviewee’s private life). This is
assumed strengthening results.
Codifying the texts is an important method for reducing datasets to equivalence concepts (Coffey &
Atkinson, 1996). This reduction enhances the possibility to analyze ‘events’ and to notice
46
similarities, differences, patterns, and structures concerning the propositions. The final datasets are
examined, during which segments and fragments in the texts, related to the propositions, are
collected and codified as ‘events’, which are aligned to specific propositions. Apart from the
answers relating to the operationalization as formulated in Table 4.5, interviews may contain
additional valuable information regarding the propositions. This information is captured in the
codification process as well. Events are analyzed and used to compare IDCs to non-IDCs. Next to
codifying fragments aligned to the propositions, fragments that are related to the (quantitative)
hypotheses are codified as well. These events are used to provide additional insights into the
empirical results in order to enrich the quantitative analyses.
47
5. Results
This chapter presents the results of this research. In line with the structure of the preceding
chapter, it is divided in two parts. The first part deals with the empirical results of the quantitative
analysis. It is based on the dataset consisting of footwear manufacturing companies in Italy and
evaluates the hypotheses introduced in paragraph 3.5.
The second part regards the qualitative part of this research. It is based on the information obtained
during the corporate and expert interviews. It addresses the propositions presented throughout
paragraph 3.5.
5.1 Quantitative results
The following sections present the results of the empirical analysis of this study. It first introduces
descriptive statistics. Hereafter, the hypothesized differences among IDCs and non-IDCs in the
dependent and explanatory variables are tested and discussed. Finally, the results of the OLS
estimation are presented and interpreted.
5.1.1 Descriptive statistics
This section offers a first insight into the dataset used throughout the empirical analysis by offering
basic descriptive statistics, which are presented in Table 5.1 below.
Variable
Mean
IDC
Standard. dev.
non-IDC
IDC
non-IDC
Minimum
IDC
Maximum
non-IDC
IDC
non-IDC
Dependent
EBITDA margin
return on assets
profit per employee
7.3317
4.7867
7.8707
5.9395
3.4779
4.7392
6.7992
9.9131
18.6330
7.5068
10.8496
15.6789
-32.8818
-66.2872
-55.2068
-36.5386
-60.8561
-88.4596
41.7230
48.3250
188.6715
32.6096
44.4298
116.0373
26.7438
72.9798
187.406
27.5444
69.2636
266.662
10.9609
15.5233
394.224
11.6008
18.9030
730.277
0.8938
7.9846
-1453.992
0.8007
0.3506
-1579.899
89.2690
98.2812
3739.033
87.7773
99.6131
6879.052
15.5502
0.9071
0.0483
0.9554
538
15.7432
0.9091
0.2584
0.9633
627
10.9338
0.2906
0.2147
0.2066
11.7847
0.2877
0.4381
0.1881
0
0
0
0
0
0
0
0
69
1
1
1
82
1
1
1
Explanatory
labor costs
specialization
leverage
Control
company age
size (% SMEs)
location (% South)
legal status
Observations
Table 5.1 Descriptive statistics
It can be observed that the sample is more or less evenly divided between IDCs (46%) and nonIDCs (56%). For the independent variables, all performance indicators’ mean values for IDCs are
higher than those for non-IDCs. Looking at the explanatory variables shows that both for labor costs
and leverage non-IDCs demonstrate higher mean values than IDCs. Remarkable is the broad range
and variance of the variable leverage. The control variables, finally, do not present any clear cut
differences. An exception is southern Italy, which demonstrates that among IDCs only 5% is located
48
in southern Italian regions, whereas for non-IDCs this mean value is remarkably higher (26%).
Moreover, it indicates that both for IDCs and non-IDCs an absolute majority of observations regards
SMEs (91%) and that for both groups around 96% of the companies are active (that is, not in
liquidation or bankruptcy).
5.1.2 Independent samples t-tests
In order to answer the research question ‘Do IDCs outperform non-IDCs?’ a t-test is used.
Table 5.2 presents the results.
T-test Dependent variables
Variable
EBITDA margin
return on assets
profit per
employee
Observations
non-IDC
IDC
non-IDC
IDC
non-IDC
IDC
627
538
627
538
627
538
(54%)
(46%)
(54%)
(46%)
(54%)
(46%)
Mean
Standard
deviation
5.939
7.331
3.477
4.786
4.739
7.870
7.506
6.799
10.849
9.913
15.678
18.633
Table 5.2 T-test dependent variables
Significance
0.001**
0.032*
0.002**
**Significant at 0.01 level (two-tailed)
* Significant at 0.05 level (two-tailed)
The results demonstrate that for any of the three considered performance indicators the mean
values of IDCs are significantly higher than those of non-IDCs. Therefore, one concludes that IDCs
outperform non-IDCs, regarding both profitability and productivity. To test for the hypothesized
differences between IDCs and non-IDCs in the explanatory variables labor costs, specialization, and
leverage another t-test is applied, whose results are presented in Table 5.3.
T-test Explanatory variables
Variable
Observations
627
non-IDC
538
IDC
627
non-IDC
specialization
538
IDC
627
non-IDC
leverage
538
IDC
Table 5.3 T-test explanatory variables
labor costs
(54%)
(46%)
(54%)
(46%)
(54%)
(46%)
Mean
27.544
26.743
69.263
72.979
266.661
187.406
Standard Significance
deviation
11.600
10.960
18.903
15.523
730.277
394.223
0.227
0.000**
0.019*
**Significant at 0.01 level (two-tailed)
* Significant at 0.05 level (two-tailed)
The results demonstrate (1) no significant difference between IDCs and non-IDCs in labor costs
(reject Hypothesis 1a), (2) a highly significantly higher mean for IDC specialization (accept
Hypothesis 2a), and (3) a significant difference for leverage. For leverage, however, the sign of the
difference is opposite to the hypothesized relationship, that is, non-IDCs demonstrate significantly
higher leverage ratios than IDCs (reject Hypothesis 3a).
49
5.1.2 Correlation matrix
The t-tests in the preceding paragraph demonstrate differences between the mean values of the
two groups (IDCs and non-IDCs). However, observed differences do not indicate any causal
relationship between the variables. Table 5.4 provides an insight into the correlation among the
variables.
The matrix shows modest correlations. Only among the dependent variables, moderate correlation
(> 0.5) is observed. This can be declared by the similarity of these variables, all being company
performance indicators and directly relying on company profitability.
Remarkable is the substantial negative correlation between EBITDA margin and specialization
(-0.41), compared to the correlation of specialization and the other two performance indicators 0.17 and 0.14). Another particularity is the difference in correlation between labor costs and profits
per employee (0.28) compared to the other two performance indicators (0.05 and 0.13). The
negative correlation (-0.28) between district and location in southern Italy confirms the
observation from the descriptive statistics.
50
Variable
Dependent
EBITDA margin
return on assets
profit / employee
Explanatory
district
labor costs
specialization
leverage
Control
age
size
location
legal status
EBITDA
margin
return
on assets
profit per
employee
1
**0.6427
**0.5780
1
**0.8145
1
**0.0740
0.0480
*-0.4107
-0.0431
0.0569
**0.1258
**-0.1749
**-0.1374
**0.0970
**0.2829
**0.1437
**-0.2221
0.0215
-0.0378
-0.0562
**0.1351
-0.0302
-0.0301
**-0.1525
**0.1488
0.0401
**-0.1277
**-0.1942
**0.1584
Table 5.4 Correlation matrix
district
labor
costs
specialization
leverage
1
-0.0308
**0.0859
-0.0171
1
**0.0985
-0.0267
1
**-0.1196
1
0.0039
-0.0035
**-0.2847
-0.0201
**0.2287
**-0.2298
**-0.2830
**0.0995
0.0330
**-0.1802
**-0.1194
0.0528
0.0261
**0.0958
-0.0443
**0.0852
age
1
-0.1738
**-0.2417
**0.0261
size
1
**0.1072
*-0.0652
location
1
-0.0286
legal
status
1
**Significant at 0.01 level (two-tailed)
* Significant at 0.05 level (two-tailed)
51
5.1.3 Regression
In order to study causal relationships, OLS regression is applied, which comprises the final part of
the empirical analysis of this research. Using Equation 4, three distinct models are tested: each
single performance indicator is used as dependent variable in a separate model. Table 5.5 presents
the results of the regression analysis.
Multiple regression analysis
Variable
Independent
Constant
constant
Explanatory
district
labor costs
specialization
specialization2
leverage
Control
age
size
location
legal status
Model summary
R2
adjusted R2
observations
1) EBITDA margin
Coefficient
Std. error
Dependent
2) return on assets
Coefficient Std. error
3) profit per employee
Coefficient
Std. error
-2.8379
2.0597
**-9.1236
3.0149
-8.9959
5.0319
**1.3413
0.0323
**0.3492
**-0.4039
**-0.0011
1.9397
0.0239
0.0505
0.0417
0.0004
0.7514
*0.0523
**0.3386
**-0.3742
**-0.0014
0.6077
0.0269
0.0767
0.0634
0.0005
*2.3647
**0.2671
**0.3629
**-0.3135
**-0.0027
1.0142
0.0449
0.1280
0.1059
0.0008
-0.0233
**-1.9463
**-1.9665
**7.7257
0.0174
0.6880
0.5771
0.9642
*-0.0653
*-2.1847
**-4.0360
**11.8676
0.0263
1.0453
0.8628
1.4620
**-0.1327
**-7.8073
*-3.5057
**8.5277
0.0440
1.7445
1.4400
2.4400
0.2147
0.2086
1165
Table 5.5 Multiple regression analysis
0.1330
0.1262
1165
0.1065
0.0995
1165
**Significant at 0.01 level (two-tailed)
* Significant at 0.05 level (two-tailed)
The results indicate that being located in an industrial district positively contributes to company
performance (significantly for Model 1 and Model 3). This is in line with the result of the t-test on
the dependent variables (Table 5.2).
In all models, labor costs have a positive influence on company performance (significantly for Model
2 and Model 3). Consequently, Hypothesis 1b is accepted. Remarkable is the difference in the
magnitude of the influence among Model 2 and Model 3 (0.05 versus 0.27).
A quadratic term (specialization2) is included. This predicts a certain optimal level of the degree of
specialization and improves the model considerably. Graphically, this optimum level can be
imagined as an inverse U-shape relationship. Apparently there exists an optimal configuration with
the respect to the degree of specialization. This outcome may relate to the optimal governance
mode, which trades off transaction costs on the one hand and the degree of asset specificity of
transactions on the other hand (compare Figure 3.3). Specialization is highly significant in all three
models and is positively related to company performance. Therefore, Hypothesis 2b is accepted.
Since the literature predicts leverage having an optimal level, a quadratic term is created, added to
the regression, and tested for. However, this quadratic term is insignificant and therefore dropped
from the model again. Leverage is highly significant for all tested models. The magnitude of the
52
coefficient, however, is negligibly negative. Hypothesis 3b is rejected accordingly.
Considering the control variables, the following can be mentioned: age has a moderately negative
effect on company performance (significantly for Model 2 and Model 3). This indicates that older
companies perform poorer than recently founded companies. This result is opposite to the
reasoning that older companies, possessing a higher degree of accumulated experience over time,
would outperform younger companies. A possible explanation could be found in leap frogging (EI 7,
2009). Size is a dichotomous variable (0: no SME; 1: SME). The results predict that smaller company
size is negatively related to company performance. In other words, larger companies are
significantly more profitable. This may be attributed to internal economies of scale. Noted should
be, however, that large companies account for only 10% of all observations, so only for a minority
of the sample. Location provides a clear signal companies from southern Italy performing
significantly poorer than companies from northern and central regions. This is in line with the
general structural economic dichotomy of Italy as discussed in paragraph 2.1. Legal status, finally,
shows coefficients with the largest magnitude and indicates that active companies perform
significantly better than companies in liquidation or bankruptcy procedures, which coincides with
common sense.
Having individually discussed every single variable, differences between the respective models are
addressed. With the highest R2 value (21%), Model 1 has the highest explanatory power of all
models on the one hand. However, Model 1 presents the lowest number of significant coefficients
(seven) on the other hand. Model 2 has intermediate explanatory power (R2: 13%) and eight
significant coefficients. Nonetheless, district is insignificant, which is considered a serious drawback
of Model 2. Model 3, finally, shows merely significant coefficients (nine, of which seven significant at
the highest α=1% confidence level). Conversely, Model 3 has the lowest explanatory power of all
considered models (R2: 11%). Concluding, every individual model presents different pros and cons.
Combining the single attributes of the different proposed models may provide a fine insight.
To distinguish for the impact of the explanatory variables on the dependent variable caused by IDC
versus non-IDC differences, interaction terms between district and all other explanatory variables
are created by multiplying district with all other explanatory variables respectively. Those are
tested in a separate model:
( )* + )% ,% + )- ,- +)%- ,% ,- + 0 + )%1 ,% ,1 . . . +)/ ,/ + $
(5)
53
The results are presented in Table 5.6. Since all interaction terms result insignificant, the analysis
demonstrates that the impact of the explanatory variables on company performance is not affected
by district, which is a noteworthy outcome. For example, the positive impact of specialization on
company performance is not significantly different between IDCs and non-IDCs.
Multiple regression analysis
Variable
Independent
Constant
constant
Explanatory
district
labor costs
specialization
specialization2
leverage
Interaction
district * labor costs
district * specialization
district * specialization2
district * leverage
Control
age
size
location
legal status
Model summary
R2
adjusted R2
observations
1) EBITDA margin
Coefficient
Std. error
Dependent
2) return on assets
Coefficient Std. error
3) profit per employee
Coefficient
Std. error
-3.5577
2.2368
**-9.5805
3.3971
-8.3789
5.6710
3.0210
*0.0545
**0.3483
**-0.4038
*-0.0009
3.4484
0.0239
0.0627
0.0534
0.0004
1.0073
0.0603
**0.3721
**-0.4174
-0.0010
5.2373
0.0363
0.0952
0.0810
0.0005
-0.6470
**0.2568
*0.3812
*-0.3474
*-0.0020
8.7429
0.0606
0.1590
0.1353
0.0009
-0.0474
-0.5294
0.3202
-0.0764
0.0345
10.8799
8.8022
0.0798
-0.0141
-4.9454
7.4785
-0.1942
0.0524
16.5237
13.3683
0.1212
0.0283
1.9338
2.768
-0.3392
0.0875
27.5840
22.3164
0.2024
-0.0250
**-1.9287
**-1.8462
**7.8341
0.0174
0.6883
0.5799
0.9647
**-0.0685
*-2.1483
**-4.1384
**11.9593
0.0265
1.0453
0.8807
1.4652
**-0.1372
**-7.7710
*-3.7549
**8.6028
0.0442
1.7450
1.4702
2.4459
0.2166
0.2077
1165
0.1366
0.1268
1165
Table 5.6 Multiple regression analysis (including interaction terms)
0.1098
0.0998
1165
**Significant at 0.01 level (two-tailed)
* Significant at 0.05 level (two-tailed)
The reader should not be confused by the insignificant coefficients of district in this model. In this
particular model with multiplicative interaction terms, the coefficient of the independent variable
district captures the impact of a company belonging to an IID or not on company performance,
keeping all other independent variables constant (and vice-versa). It does not reflect the impact of
district on company performance in general. Because interactive relationships imply that the effect
of district on company performance varies according to the level of the other independent variables,
the notion of the impact of district on company performance is in fact a meaningless one
(Braumoeller, 2004).
54
Table 5.7 provides an overview of the quantitative results. Reviewing the quantitative results, one
can state that (1) IDCs outperform their non-IDC counterparts, (2) only one of the explanatory
variables (specialization) is accepted in the empirical analysis. Labor costs are positively related to
performance, but labor costs do not differ significantly between IDCs and non-IDCs. Remarkable is
the observed performance differential between companies from southern Italy (lower
performance) and the rest of Italy (higher performance).
Feature
Statement
Labor market
Hypothesis (1a)
Labor costs are higher for IDCs compared to non-IDCs
Hypothesis (1b
Labor costs are positively related to IDC performance
Specialization
Hypothesis (2a)
IDCs are more specialized compared to non-IDCs
Hypothesis (2b)
Specialization is positively related to IDC performance
Financial intermediaries
Hypothesis (3a)
IDCs have higher leverage ratios compared to non-IDCs
Hypothesis (3b)
Leverage ratio is positively related to IDC performance
Table 5.7 Quantitative results overview
Result
X
√
√
√
X
X
These preliminary results might indicate that there are other factors into play. The next paragraph
considers a number of more qualitative aspects related to IIDs, which possibly contribute to IDCs’
performance.
5.2 Qualitative results
The following sections present the results of the qualitative analysis of this study. The results
provide deeper insight into the different propositions concerning the features of IIDs presented
throughout paragraph 3.5.
5.2.1 Labor market
The production process in the footwear sector consists of several phases. The respondents indicate
that certain phases of the manufacturing process require specialized employees, who possess
extensive and specific experience in working with particular machines accountable for key phases
of the production process. Moreover, they are responsible for quality management. The use of
specialized employees on these key positions is believed to be crucial for ensuring the Italian
quality standard (Table 4.4, corporate interview (CI) number 3, 2009). The indispensable role of the
specialist creates bargaining power for these specialized suppliers of labor, which can result in
poaching, a process driving up wages. The quantitative analyzes, however, do not provide evidence
for a significant wage difference between IDCs and non-IDCs.
There are, nevertheless, other ways of attracting specialized personnel. One company offered
specialized personnel flexible working hours to attract them from other industrial sectors (CI 17,
2009). The presence of this indirect poaching confirms the theory. The degree of poaching depends
on the local availability of personnel.
The existence of unregistered (and possibly unpaid) labor by family members as suggested by
Hadjimichalis (2006) is denied by the interviewees (CI 2; 5; 8; 18; 22, 2009). However, ‘…there are
55
those companies in southern Italy, which don’t pay their workforce so they spoil the prices and ruin the
market. They have a lot of workers paid in black... … Moreover, we have to stick to many standards
(security, working conditions etc.) and are controlled regularly. We have to pay for these inspections
ourselves. Those unauthorized workshops don’t incur such costs. It is unfair competition! ’(CI 21, 2009).
So the presence of unregistered labor and tax evasion is denied by interviewees on the one hand
and the same interviewees accuse other, southern Italian, companies of unfair competition on the
other hand. In the end, the question remains: if verifiable at all, would this phenomenon differ
significantly between IDCs and non-IDCS?
Generally, interviewees mention the lack of experienced personnel and the trend of youngsters not
being interested anymore in working in footwear manufacturing. This is a serious threat for the
continuity of the sector, as generational transition is considered crucial for maintaining the artisan
skills accumulated throughout decades as a public good within the boundaries of IIDs’ community
(CI 5; EI 7, 2009).
Companies obtain their employees through various channels. Some employees present themselves
directly at the company, for example, after their former employer ceased its activities. All
companies use word of mouth as dominant strategy for finding suitable personnel. This confirms
the importance of the local community network as important IID feature. Next to informal channels,
a minority of the interviewees uses formal services of employment agencies (CI 4; 16; 18; 22, 2009).
The time span needed for employing new suitable personnel depends on the local availability of the
required personnel and differs in the interviews to from three days (CI 22, 2009)up to one month
(CI 17, 2009).
The majority of companies’ employees live within a 20 minute radius from their working place, as
Italian employees usually prefer having lunch at home. This limits the geographical range within
which companies can recruit. Therefore, IDCs would benefit from the large number of potential
employers present in IIDs. The information obtained in the interviews, however, does not
unequivocally confirm Proposition 1.
5.2.2 Industrial organization
The following two subsections present the results of the qualitative analysis of two topics on
industrial organization: (1) vertical inter-firm cooperation and (2) horizontal inter-firm
competition.
5.2.2.1 Vertical inter-firm cooperation
The quantitative analysis provides evidence that IDCs are more specialized than non-IDCs
(acceptation of Hypothesis 2a). This implies that for transactions IDCs make relatively more use of
normal cooperation than vertically integrated companies. Transaction costs theory prescribes
normal cooperation as optimal governance mode, if transactions are characterized by relatively
high levels of asset specificity, high frequency, and high uncertainty.
An additional explanation for higher specialization of IDCs could simply relate to the larger
presence of active companies and potential suppliers in a geographically limited area. All
interviewees acknowledge the advantage of IIDs regarding this feature (CI 4; 15; 18; 19; EI 1, 2009).
The bargaining power of suppliers is increased by their reputation of delivering high quality and at
the same time decreased by the high number of suppliers. This last point is more applicable to the
56
situation in IIDs: the substantial amount of suppliers increases sellers’ bargaining power and
therefore offers possible profits.
Interviewees indicate that investments in machinery are substantial and occupy a considerable part
of companies’ working capital (CI 10; 12; 18, 2009). These are long-term investments, whose
repayment horizon is long-drawn accordingly. The extent of return on investment requires a stable
customer demand, if the increased production capacity is to be exploited optimally. If customer
relationships were short-term focused, substantial long-term investments would become dedicated
assets, exposing companies to a high degree of uncertainty. Deployed in their second best
possibility, dedicated assets lose considerable value. Nevertheless, IDCs invest in such assets.
Companies remain using normal cooperation, despite uncertainty and risk of higher transaction
costs. Since IDCs outperform non-IDCs, there should be an explanation for this observation.
All interviewed companies indicate that they possess long-term supplier relationships. These
relationships last for several years (CI 15, 2009) and are predominantly reassessed periodically in
verbal, non-formalized agreements (CI 1, 2009). The use of long-term relationships decreases
uncertainty and enables long-term investments in dedicated assets. This possibly explains the
observed high degree of specialization among IDCs.
Using verbal agreements makes part of local customs and decreases transaction costs, as the
creation of written contracts requires time, effort, and costly legal services. So despite the high
frequency of transactions, integrating transactions is unnecessary as transaction costs remain
relatively low, due to the absent need for preparing formal contracts. The presence of verbal
agreements indicates the existence of trust between parties engaging in transactions. Normally
spoken, however, written contracts would provide higher legal insurance. Companies remain loyal
to each other: interviewees indicate that companies prefer remaining with existing suppliers even
in case lower prices are offered by alternative suppliers (CI 10, 2009).
This loyalty applies during economic downturns as well. In periods of decreasing demand,
companies reduce production and, consequently, orders to their suppliers. Instead of reducing the
total number of suppliers by ceasing the relationship with one or more suppliers, companies
consciously evenly reduce the total number of orders divided over the entire group of existing
suppliers. One of the reasons forms the insurance for quality, as companies trust their suppliers in
delivering according commonly agreed quality standards (CI 9, 2009).
Both seller and buyer of supplies have their reputation to be secured. One interviewed company
even developed a system to quantify reputation and monitor performance of its different suppliers
(CI 9, 2009) and several interviewees mention on the spot inspections of their subcontractors (CI 8;
18, 2009). Reputation appears an important mechanism in reducing uncertainty and ensures the
quality of delivered goods.
Information with respect to cooperation is mixed and varies from ‘According to me we (red.
footwear producers from the Marche region) aren’t very intelligent from our cultural background. In
my vision there should be cooperation among us to affront problems we all have to deal with anyway.
However, we don’t interchange ideas or opinions; we are all a bit jealous, you know. According to me,
this is a pity.’ (CI 10, 2009) to ‘We have a good relationship with them (red. competitors) and it (red.
57
cooperation) is more intense compared to competitors located further away... … We exchange raw
materials and we use each other’s machines in case of idle capacity’ (CI 11, 2009).
Nonetheless, such intense form of cooperation does not take place solely in IIDs. Non-IDCs make
use of long-term contracts (CI 19, 2009) and verbal agreements as well (CI 15, 2009). Moreover,
there appears to be no clear difference in the usage degree of normal cooperation between IDCs
and non-IDCs. Apparently, long-term buyer-seller relationships and verbal agreements (instead of
written contracts) belong to Italian business customs in general rather than to local customs in IIDs
exclusively.
Concluding, there is no indication for accepting Proposition 2a.
5.2.2.2 Horizontal inter-firm competition
The higher number of footwear companies in footwear IIDs enables companies cooperating in an
efficient way. This is the case for companies with substitutable activities. However, the impact of a
higher number of competing companies is unclear.
All interviewed companies have competitors based in Italy. IDCs have several competitors located
within the same IID (CI 7, 2009). Non-IDCs have competitors as well; however, those competitors
tend to be less strongly collocated (CI 15, 2009). Non-IDCs do not directly perceive the number of
competitors in IIDs higher than outside IIDs, as they do not regard competition as geographically
bounded (CI 7; 15, 2009). Following this perspective, there may not be a difference in the number
of competitors.
However, the relationship between competitors appears to be different between IDCs and nonIDCs. Competing non-IDCs do not collaborate in whatever way and only meet occasionally during
trade fairs (CI 15; 16, 2009). Some IDCs have a similar distant relationship to their competitors (CI
3; 7; 10, 2009). Interviewees’ answers with respect to the exchange of information present
ambiguity. They differ from no exchange at all: ‘…this (red.: absence of knowledge exchange) is a big
defect which we, Marchigiani (red.: inhabitants of the Marche region), have: there is little cooperation
among firms. We could do so many things, but we never did. During the years there have been many
attempts, as well initiated by trade associations, but it never worked out’ (CI 9, 2009), no exchange
with direct competitors (CI 17) to exchange of information exclusively with befriended competitors
(CI 8; 13, 2009), to free exchange of materials (CI 14, 2009), exchange of materials, advice, and
personal interaction (CI 18), and mutual support and exchange of advice (CI 2, 2009). Nevertheless,
IDCs acknowledge the possible benefits of collaborating with competitors (CI 7; 10, 2009)
Other IDCs are actively engaged in forms of collaboration with collocated competitors, trying to
reach those benefits (CI 2; 4; 5, 2009). Examples are a jointly marketing campaign (CI 4, 2009),
collective bargaining with dispatchers, and cooperation on fiscal consultancy issues (CI 5, 2009).
So differences appear to exist between IDCs and non-IDCs in their relationships with competitors.
The existence of competitors is acknowledged, but not directly perceived as disadvantageous by
IDCs. Competition stimulates competing companies in outperforming competition in a continuous
strive for efficiency and innovation: ‘Cooperation and competition -the will of outdoing the otheralways come together. We cooperate where we can. This creates stimuli.’ (CI 5, 2009). This is a
possible explanation for increasing IDC performance, confirmed by the theory.
58
Summarizing, interviewees do not perceive a substantial difference between IDCs and non-IDCs
with respect to the number of competitors, as competitors are located throughout the whole of Italy
and competition is regarded becoming ever more global. There seems, however, a difference
between IDCs and non-IDCs in terms of the relationship with their competitors, as some IDCs have
started nascent collaborations with competitors. This is not directly in line with Proposition 2b,
which predicts more intense horizontal competition between IDCs.
5.2.3 Institutions
Several trade organizations are active in the footwear sector. Trade organizations officially
differentiate their services towards companies primarily depending on company size. As can be
observed in Figure 2.2, for example, Confindustria is primarily focused on medium-sized
companies, while CNA on small-sized companies, or artisans (CI 18, 2009). Trade organizations
therefore do not seem to be geographically bounded and to specifically aim their services at IDCs
(Table 4.4, expert interview (EI) 6, 2009), which rejects Proposition 3a. The observation that nonIDCs affiliate to trade organizations as well (CIs 16; 19, 2009) rejects Proposition 3b. Apart from
size, there might exist other reasons for companies affiliating to certain trade organizations, as
interviewees indicate that trade organizations are politically oriented (CI 19, 2009).
Membership of trade organizations generally requires a fee, so companies have to evaluate if the
benefits derived from membership outweigh its costs. Services offered by trade organizations
mainly consist of (1) financial services (consulting and risk sharing), (2) trainings, (3) collective
lobbying activities, and (4) the organization of trade fairs.
An example of the first category is payroll services. However, these are not used by all members, as
some companies hire an external accountant fulfilling this task (CI 18, 2009). Moreover, trade
organizations provide consultancy services related to ‘continuously changing legislation in Italy, which
is incomprehensible to entrepreneurs’ (CI 16, 2009).
Trainings are aimed at the development of personnel and consist of, for example, courses in
marketing management, internationalization, and foreign languages. Some respondents have
experience in this kind of courses (CI 19, 2009). Offering trainings relates to the third task of
institutions as proposed in paragraph 3.5.2 (Dei Ottati, 2002).
Lobbying activities aim at strengthening political support for the sector and at favorably influencing
industrial policy. One example is the lobby for exclusivity of the ‘made in Italy’ label. This label is
considered a way of diversifying Italian products from competing goods from low labor costs
countries. Since consumers generally are willing to pay a premium for products made in Italy
(because many consumers associate Italian products to modern design, style, and high quality),
protecting its image and countering counterfeiting are considered important industrial policy
issues. Via the membership of trade organizations, individual companies combine their forces in
collective lobbying activities on the national and European level (CI 9, 2009). Trade organizations
organize meetings for active members to discuss lobbying activities, which relates to the first task
of institutions as proposed by Dei Ottati (2002): determination of a long-term vision. However, the
efficiency of such meetings is questioned by the majority of interviewed members. Since trade
organizations are managed by representatives from the footwear sector itself, it appears difficult to
achieve consensus, as every actor primarily tries to fulfill his own particular interests (CI 10, 2009).
National and international trade fairs are organized by ANCI, the national Italian footwear
59
association. Attending trade fairs is seen as an opportunity for companies for brand exposure,
networking, and attracting new customers. Participating in trade fairs is perceived costly and
therefore more relevant for relatively larger companies having their own brands.
Since both IDCs and non-IDCs make use of services provided by trade organizations Proposition 3c
is rejected.
Perceived efficiency of trade organizations differs among interviewees. However, it seems to
depend more on the degree to which companies actively engage in trade organizations themselves
rather than on company location (inside or outside IIDs). Those interviewees actively participating
in trade organizations express a more positive attitude (CI 5, 2009) compared to passive member
interviewees (CI 14, 2009). Generally spoken, none of the interviewees expressed extreme
satisfaction by trade organizations: either the entire gamma of offered services is not used complete
(yet), or existing services have potential for improvement. Therefore, Proposition 3d is rejected.
Summarizing, it can be said that apparently a difference between IDCs and non-IDCs regarding the
perceived efficiency of trade organizations does not exist. Trade organizations fulfill some tasks as
proposed by Dei Ottati (2002). However, their beneficial impact does not seem to differ between
IDCs and non-IDCs. The impression obtained during the corporate interviews suggests that
perceived benefits derived from membership do not depend on company location, but rather on the
willingness to actively participate and engage in trade organizations and their activities.
5.2.4 Financial intermediaries
The presence of local banks forms a distinctive feature of the Italian economy in general and of
industrial districts in particular. All corporate interviewees are aware of the existence of local
banks and an absolute majority uses services provided by local banks. A number of the
interviewees use local, regional, as well as national banks (CI 8; 9; 14; 19, 2009). Relationships with
local banks are described as intense and face-to-face or telephone interaction between bank and
company is described as frequent (daily/weekly), even since the introduction of online banking (CI
5; 8, 2009).
An interviewed local banker indicates that the main competitive advantage of local banks compared
to national banks consists of better knowledge of the client’s situation leading to a different client
relationship strongly based on mutual trust, allowing generally for quick decision taking especially
in case a company needs liquidity immediately (EI 4, 2009). These advantages are confirmed by the
literature. Given that these advantages reduce local banks’ monitoring costs, this may induce local
banks providing relatively more credit given companies’ equity level, leading to higher leverage.
However, this hypothesis is rejected by the empirical analysis.
Corporate interviewees mention the proximity and personal relationship of a local bank compared
to a national bank (CI 18; 22, 2009). To illustrate: ‘We have both local as national banks. We work
much better with the local ones, because they are closer to our reality. The national banks and their
decision making centers are far away and above all they aren’t attuned to the problems of small
enterprises. In the Marche region we have financial institutions which -after series of mergers and
acquisitions- have become quite robust and of considerable size, however, keeping a local structure and
identity, which we like’ (CI 5, 2009). However, ‘In the end it depends on what kind of relationship you
are able to establish with the banker; it depends strongly on the capability of the persons how to manage
60
the relationship with the bank. I am convinced that the relationship between the director of the local
bank branch and the company is very valuable. This can be both with national and local banks’ (CI 10,
2009). These quotes show that the range of answers is broad and unequivocal.
EI 4 (2009) reveals that local banks experience lower economies of scale compared to national
banks and that national banks are increasingly active and acting locally as well, adopting strategies
similar to local banks (tailored customer service based on thorough understanding of the
customer’s reality). National banks do so by acquiring local branches or opening up local
subsidiaries, which allows for benefiting both from scale economies and scope by the adopted
localized strategy.
Summarizing, the information obtained during the interviews clarifies the properties and perceived
advantages of using the services of local banks. However, the information does not give indications
that IDCs make more often use of the services of local banks compared to non-IDCs. Accordingly,
Proposition 4 is rejected.
Table 5.7 provides an overview of the results of the qualitative analysis. It shows that the entire set
of propositions is rejected based on the information obtained during the corporate and expert
interviews.
Feature
Labor market
Statement
Proposition (1)
IDCs obtain faster suitable employees than non-IDCs
Result
X
Industrial organization
Proposition (2a)
Proposition (2b)
IDCs exhibit a higher degree of vertical inter-firm cooperation than non-IDCs
IDCs exhibit more intense horizontal inter-firm competition than non-IDCs
X
X
Institutions’ services are specifically targeted towards IDCs
IDCs are more often associated to institutions compared to non-IDCs
IDCs make more use of services provided by institutions
IDCs are more satisfied by the services provided by institutions
X
X
X
X
Institutions
Proposition (3a)
Proposition (3b)
Proposition (3c)
Proposition (3d)
Financial intermediaries
Proposition (4)
IDCs make more often use of local banks compared to non-IDCs
X
Table 5.8 Qualitative results overview
Even if all propositions are rejected, the qualitative analysis does propose some interesting
outcomes: (1) an indication for alternative labor poaching by offering secondary benefits, (2) no
indication for the presence of unregistered labor, (3) a wider range of possibilities for IDCs with
respect to vertical inter-firm cooperation, (4) a different perception of horizontal inter-firm
cooperation between IDCs and non-IDCs, (5) more appreciation for the role and impact of trade
associations with high levels of participation in those associations, and (6) a diminishing difference
between local banks on the one hand and national banks on the other hand.
61
5.3 Summary
This chapter presented the results of this research. Both quantitative and qualitative analyses are
deployed in order to test the hypotheses and propositions. Table 5.8 summarizes the research
results.
Feature
Labor market
Statement
Result
Proposition (1)
Hypothesis (1a)
Hypothesis (1b
IDCs obtain faster suitable employees than non-IDCs
Labor costs are higher for IDCs compared to non-IDCs
Labor costs are positively related to IDC performance
X
X
√
IDCs are more specialized compared to non-IDCs
Specialization is positively related to IDC performance
√
√
Specialization
Hypothesis (2a)
Hypothesis (2b)
Industrial organization
Proposition (2a)
Proposition (2b)
IDCs exhibit a higher degree of vertical inter-firm cooperation than non-IDCs
IDCs exhibit more intense horizontal inter-firm competition than non-IDCs
X
X
Institutions’ services are specifically targeted towards IDCs
IDCs are more often associated to institutions compared to non-IDCs
IDCs make more use of services provided by institutions
IDCs are more satisfied by the services provided by institutions
X
X
X
X
Institutions
Proposition (3a)
Proposition (3b)
Proposition (3c)
Proposition (3d)
Financial intermediaries
Hypothesis (3a)
Hypothesis (3b)
Proposition (4)
IDCs have higher leverage ratios compared to non-IDCs
Leverage ratio is positively related to IDC performance
IDCs make more often use of local banks compared to non-IDCs
X
X
X
Table 5.9 Results overview
These results indicate that for the majority of features under study no clear-cut differences between
IDCs and non-IDCs are found. The effect of such potential differences on company performance is
acknowledged, even though this research provides evidence for a performance differential between
IDCs and non-IDCs. The accepted hypotheses indicate the following:
•
•
•
labor costs have a negligible (coefficient 0.05, Model 2) to moderate (coefficient 0.26, Model
3) positive effect on company performance (Hypotheses 1b). However, no significant
difference in labor costs is determined between IDCs and non-IDCs (Hypotheses 1a);
the existence of unregistered labor by family members and family members exceeding
regular working hours are denied by the interviewees. However, unfair competition due to
other companies not respecting regulations is claimed for;
IDCs are more specialized than non-IDCs (Hypothesis 2a) and specialization positively
affects company performance (all Models demonstrate highly significant positive
62
•
•
•
coefficients around 0.35). Even though specialization is positively related to IDC
performance, this positive relationship cannot be significantly explained by IDC versus nonIDC differences as shown by the highly insignificant interaction terms between
specialization and district;
Institutions appear to have no distinct effect. The perceived benefits seem to be more
related to the degree of participation than localization;
In contrast to the hypothesis, leverage demonstrates higher value for non-IDCs than IDCs.
Next to that is the magnitude of the effect of leverage on company performance negligible;
Geographical differences in Italy’s economic structure cause significant performance
differentials: companies from southern regions clearly underperform companies located in
other regions.
63
6. Conclusion
This conclusion synthesizes the research question and results of this study. More detailed answers
are profoundly presented and discussed throughout Chapter 5. This conclusion is followed by a
critical assessment of the chosen research design and data analysis in order to provide a complete
view on the validity of this research and, consequently, of its conclusions. Finally, it leads to a
number of directions for further research.
The use of different data sources improves the insight into IIDs from the footwear sector. The
combined use of qualitative as well as quantitative data supplements and enables providing a
deeper understanding of the performance increasing features of IIDs. The difficulties related to
finding suitable data for various, partially intangible, features make the use of multiple data sources
even more important. By means of the quantitative and qualitative analyses, an answer on the main
research question is presented.
Do companies located in industrial districts outperform their non-district counterparts and why?
The results of the quantitative analysis reveal that, firstly, IDCs differ from non-IDCs with respect to
company performance. IDCs outperform their non-district counterparts on all three company
performance indicators considered by this research. This supports the conjecture of the existence
of a district effect. The explanation for this effect can be found in several characterizing features of
IIDs. Secondly, two features under study (labor costs and specialization) demonstrate a positive
relationship with company performance. Notwithstanding the positive effect of labor costs and
specialization on company performance, it does not differ significantly between IDCs and non-IDCs.
Microeconomic theory predicts a positive relationship between labor costs, labor productivity, and,
finally, profitability. Nevertheless, the empirical analysis does not find direct supporting evidence of
higher wages for IDCs. IDCs provide employees additional secondary benefits, which could explain
higher labor productivity, though.
The industrial organization of IIDs is characterized by a relatively high degree of phase
specialization. IDCs are more specialized than non-IDCs. This higher degree of specialization results
in more efficient forms of vertical inter-firm cooperation, which is company performance
enhancing. A final observation from the quantitative analysis refers to the significant performance
differential between southern regions and the rest of Italy.
The information obtained in the corporate interviews provides an equivocal image on horizontal
inter-firm competition. The degree of competition is perceived to be similar, especially since new
technologies and means of communication have brought competition to the global playing field.
However, whereas IDCs experience competition as beneficial and stimulating, non-IDCs do not
regard competition as advantageous but rather as a threat. There are no differences perceived in
the efforts of institutions, which therefore does not offer an explanation for higher IDCs
performance. There is no difference observed with respect to the role of financial intermediaries.
This does not form an explanation for IDCs performance either.
64
6.1 Limitations
The results of the quantitative analysis provide evidence for a significant ‘district effect’, that is,
showing significantly higher performance among IDCs compared to non-IDCs. The observed
differences in company performance between IDCs and non-IDCs cannot be attributed to the
explanatory variables considered by this study, however. Other characteristics of IIDs can influence
IDCs’ performance as well. Knowledge spillovers, for example, which are recognized as a
distinguishing characteristic of IIDs contributing to company performance, are not subject of
current study. Nevertheless, a respondent mentions those mechanisms as a main reason for the
current favorable position of their company within the footwear sector (CI 7, 2009).
The quantitative data refer to 2007, a period of economic stability and relative prosperity for the
footwear sector. However, the qualitative dataset is obtained during a period of a severe global
economic crisis, with a real negative impact on the footwear sector, affecting global trade and
declining trade volumes. Since IIDs partially export their products to international markets
(D’Arpizio, 2008, Centro Studi Confindustria Marche, 2008), interviewed companies are expected to
be exposed to the economic crisis, which could have influenced the information gathering. Despite
the efforts of the interviewers to relate the questions to the relevant reference period (that is,
before the economic crisis in 2007), it is possible that the perception of the interviewees could still
be affected by the downturn in the footwear sector at the moment of the actual interview.
What should be acknowledged about the sample selection through Amadeus, is that technically
spoken one cannot even speak of a ‘sample’. In this context, namely, the selection criteria of the
companies are not rigorously determined by Amadeus. Next to that, Amadeus’ coverage decreases
with size, which might indicate an underrepresentation of small companies. It is crucial to take this
possible bias into consideration, as this study’s particular focus is industrial districts, which are
characterized by a strong presence of SMEs.
The Sforzi-ISTAT classification method is used for identifying IIDs. Although widely applied,
originally it is not intended for distinguishing IIDs (but for mapping commuting movements), which
leads to a number of drawbacks. For example, it is possible that an LLMA is specialized in two
different sectors and, therefore, should belong to two separate districts. One LLMA belonging to
multiple districts, however, is ruled out by the Sforzi-ISTAT criterion. Consequently, it is possible
that the boundaries of IIDs are not always in line with observations from the field. Moreover, the
last time that IIDs are determined by means of the Sforzi-ISTAT classification has been in 2005
using 2001 census data. It is plausible that the situation in LLMAs has changed in the mean time.
Finally, the most serious drawback of the Sforzi-ISTAT classification is that ‘it does not account for
two of the most essential characteristics of industrial districts described by the Florentine school, that is,
the organizational and the cultural dimension’ (Boschma, 2003).
Next to the limitations of the quantitative analysis, the qualitative analysis presents a number of
restrictions as well. First, all interviews are held with Italian respondents. The interviews are
conducted either in Italian or English, depending on interviewees’ preference (that is, the vast
majority in Italian). The use of the Italian language increases the possibilities of this study, since
only a minority of the respondents is proficient in English. However, as Italian is not the native
language of the interviewers, it is plausible that some information is misinterpreted or not
65
transmitted at all.
Secondly, respondents are partially contacted through introduction by two local contacts. The use
of local contacts facilitated arranging appointments with company representatives considerably.
However, the use of the network of the contacts might result in a biased sample selection of
interviewees, unrepresentative for the entire population.
Finally, the limited number of interviews does not permit empirical testing of the propositions. This
means that inferences drawn based on the information obtained during the interviews have only
limited interpretative validity.
6.2 Recommendations
Since this is a case study, the results are only applicable to the sector under study. There is no
pretence of generality to this regard. Paniccia (2000) mentions that a sector’s technology can play
an important role in districts’ characteristics. Further research could address other industrial
sectors as well, including more technology intensive IIDs (for example: biotechnology, ICT, or
nanotechnology).
Even within the footwear sector, distinctions exist between the different market segments in which
IIDs operate. IDCs active in lower market segments, focusing more on quantity than quality, may
have different characteristics compared to IDCs active in higher market segments. As EI 1, 2009
states: ‘Barletta (red. see Figure 2.1) is completely different from Fermo-Macerata in terms of quality of
the footwear and the target market. One could compare both companies from Barletta with Fermo with
isolated companies in order to make a more general statement.’ Therefore, enlarging the qualitative
analysis deployed in present research to several other footwear IIDs is welcomed.
Enlarging the number of interviews would increase the explanatory strength of the results as well,
since the present number is insufficient for empirical testing of the propositions. Future research
could use the existing questionnaire and increase the number of interviews in order to empirically
test the propositions.
Since this study’s principal outcome relates to the degree of specialization (or vertical
disintegration put more precisely) and its impact on company performance, it is suggested using
the other commonly applied operationalization for specialization as well in future research
(calculating the vertical ratio approximating the percentage of total product which is part of a
company’s vertical chain) and considering to what extent results may differ when adopting this
alternative operationalization.
This study focuses on the nation Italy. The direct influence of European institutions on IDCs’
performance has not been taken into consideration, despite the substantial attempts of initiatives
by European institutions aimed at improving the Italian economy. Future research could address
this issue. The same applies to the effects of globalization and to the degree of internationalization:
several companies have (partially) outsourced labor intensive activities to low cost countries. This
has changed the configuration of companies in IIDs and could, therefore, affect the explanatory
features of IDCs’ performance.
66
Although this research considers inter-firm cooperation, other mechanisms for knowledge
transmittance exist at the level of industrial districts. An interesting topic for future research is the
role of knowledge spillovers and their relation to company performance. At first, one could think of
studying company age and its effect on IDC performance from a multi-generational perspective. In
such a set-up, one could address ownership succession (especially among family owned companies)
and the role of knowledge being passed on to future generations thereby looking at differences
between IDCs and non-IDCs (Cuccuelli & Micucci, 2008).
Secondly, knowledge spillovers can be approached from a spatial perspective: one could study the
degree of labor mobility and its influence on IDC performance (Breschi & Lissoni, 2003). Such a
study could address localized knowledge spillovers in a context of entrepreneurs originating from
an industrial district area moving, for example because of marriage, to another area.
Another interesting topic for future research in industrial districts, which gained considerable
attention recently, is the concept of ‘business groups’ (Cainelli et al, 2004; Cainelli & Iacobucci,
2009a). This stream discusses the existence of networks consisting of ‘leading companies’ at the
head of a business group and their subcontractors. Business groups are defined as ‘sets of
companies that are legally distinct, but which are owned and controlled by the same person(s)’ and
considered the appropriate unit of analysis when studying company organization and behavior
(Cainelli & Iacobucci, 2009b). In contrast to the relatively more informal nature of inter-firm
relationships in IIDs as sketched by prevailing literature, those leading companies of a business
group are assumed having more formalized hierarchical relationships with subcontractors through
direct ownership ties (a leading company possessing a majority share in the underlying company)
or other forms of stable inter-firm connections (alliances, franchising, joint-ventures, etc.). Future
research could address the impact of this particular organizational configuration and its impact on
performance, thereby drawing a comparison between IDCs and non-IDCs.
6.3 Final consideration
After having thoroughly investigated industrial districts and having executed this particular
research on the performance of companies from the Italian footwear industry and the effect of
being located in an industrial district on company performance, it is worthwhile taking a final
moment to zoom out and review this research and its implications again from a macro view. This
allow for asking the question: ‘What do we know now and why is that relevant?’
Concluding, this research has provided indications that the concept of IIDs is still not completely
understood, despite the substantial existing literature body. The research provides evidence that
location matters and being located in IIDs is beneficial for company performance. However, not all
mechanisms active in IIDs have been identified or incorporated in present study. Future research
can address additional mechanisms in order to enhance IID understanding. Understanding IIDs
better is of interest to all stakeholders involved -ranging from policy makers and academic
researchers, to employees and decision making mangers- and relevant both from a scientific and
from a societal viewpoint.
67
7. Acknowledgements
The authors wish to sincerely thank Prof. Dr. Ron Boschma and Dr. Sandra Phlippen for their
valuable support, professional suggestions, and their precious time dedicated during the design,
progress, and completion of this thesis.
Moreover, the availability and contributions of all interviewees who participated in this scientific
research are highly recognized. The authors are particularly grateful to Marly Janssen and Antonio
Muzi for introducing the authors to their personal networks in the footwear sector in the FermanoMaceratese area. For providing data and information the support of Luigi Cannari from Banca
d’Italia is acknowledged.
Finally, the support and suggestions of a number of referees, including Maikel Kishna and Arjan
Tiessen, are strongly appreciated.
68
8. References
8.1 Literature
-
Adelman, M., 1995. Concept and statistical measurement of vertical integration, in Adelman,
M., 1955.
-
Adelman, M., 1955. Business concentration and price policy, Princeton University Press, New
Jersey
-
Alberti, F., 2003. Re-writing industrial districts. A challenge to the taken for granted
grammar. The 17th Nordic conference on Business Studies, Scandinavian Academy of
Management, Reykjavik, Iceland.
-
Alessandrini, P., Presbitero, P. & Zazzaro, A., 2008. Banche e imprese nei distretti industriali.
University of Ancona, Quaderni di ricerca n. 309.
-
Aureli, S., Ciambotti, M. & Salvatori, F., 2008. Interpretative analysis of the Italian shoemaking
industry: some empirical evidence from the Marche region. Paper presented at XXII Rent
Conference ‘Entrepreneurship as an Engine for Regional Development’, Covilhã, Portugal.
-
Baarda, D., Goede, de M., & Teunissen, J., 2005. Basisboek kwalitatief onderzoek: Handleiding
voor het opzetten en uitvoeren van kalitatief onderzoek. Tweede Druk, Wolters-Noordhoff
bv, Groningen/Houten.
-
Baffigi, A., Pagnini, M. & Quintiliani, F., 2000. Localismo bancario e distretti industriali:
assetto dei mercati del credito e finanziamento degli investimenti, in Signorini, L.F., (eds.)
2000. Lo sviluppo locale Un’indagine della Banca d’Italia sui distretti industriali.p.237-256.
-
Becattini, G., 1990. The Marshallian industrial district as a socio-economic notion, in Pyke,
F., Becattini, G. & Sengenberger, W., (eds.) 1990. Industrial districts and inter-firm cooperation in Italy p. 10-19.
-
Becattini, G., Bellandi, M., Dei Ottati, G. & Sforzi, F., 2003. From industrial district to local
development: an itinerary of research, Edward Pelger, Northampton USA
-
Becattini G. & Dei Ottati, G., 2006. The performance of the Italian industrial districts and
large enterprise areas in the 1990s. European Planning Studies, 14 (8) p. 1139-1162.
-
Belussi, F. & Pilotti, L., 2002. Knowledge creation, learning and innovation in Italian
Industrial districts. Geografiska Annaler. Series B, Human geography, 84 (2) p. 125-139.
69
-
Boari, C., 2001. Industrial clusters, focal firms, and economic dynamism: A perspective from
Italy.
Boschma, R., 2003. Review of From industrial district to local development An itinerary of
research, Utrecht University
-
Boschma R. & ter Wal A., 2005. Knowledge networks and innovative performance in an
industrial district: The Case of Footwear District in the South of Italy. Industry and
Innovation, 14 (2) p. 177-199.
-
Braumoeller, B., 2004. Hypothesis testing and multiplicative interaction terms.
International Organization, 58 p. 807 - 820.
-
Brealy, R., Myers, S. & Marcus, A., 2004. Fundamentals of Corporate Finance. Fifth edition,
McGraw Hill companies, New York.
Breschi, S. & Lissoni, F., 2003. Mobility and social networks: Localized knowledge spillovers
revisited. Centro di Ricerca sui Processi di Innovazione e Internazionalizzazione WP n.142.
-
-
Breschi, S., Lissoni, F. & Montobbio, F., 2005. The geography of knowledge spillovers:
Conceptual issues and measurement problems, in Breschi, S. & Malerba, F., 2005. Clusters,
networks, and innovation
-
Breschi, S. & Malerba F, 2005. Clusters, networks, and innovation, Oxford University Press,
Oxford
-
Bronzini, R., 2000. Sistemi produttivi locali e commercio estero: un’analisi territoriale delle
esportazioni italiane, in Signorini, L.F., 2000. p. 101-122.
-
Brunello, G., Graziano, C. & Parigi, B., 1999. Ownership or performance: what determines
board of directors’ turnover in Italy?
-
Brusco, S., 1986. Small firms and industrial districts: the experience of Italy. In Keeble, D. &
Wever, E., (eds.) 1986. New Firms and Regional development in Europe. Croom Helm,
London.
-
Brusco, S., 1990. The idea of industrial district. Its genesis, in Pyke, F., Becattini, G. &
Sengenberger, W., (eds.) 1990. Industrial districts and inter-firm co-operaiton in Italy,
International Institute for Labour Studies, Geneva, p. 10-19.
-
Cainelli, G., Iacobucci, D. & Morganti, E., 2004. Spatial agglomeration and business groups:
new evidence from Italian industrial districts, Working paper.
-
Cainelli, G. & Iacobucci, D., 2009a. Do agglomeration and technology affect vertical
integration? Evidence from Italian business groups, Centro Interuniversitario di Economia
70
Applicata alle Politiche per l’Industria, lo Sviluppo Locale e l’Internazionalizzazione,
Working paper 3/2009.
-
Cainelli, G. & Iacobucci, D., 2009b. Business groups and the boundaries of the firm, Working
paper
-
Cannari, L. & Signorini, L.F., 2000. Nuovi strumenti per la classificazione dei sistemi locali,
in Signorini, L.F., (eds.) 2000. Lo sviluppo locale Un’indagine della Banca d’Italia sui distretti
industriali.
-
Capecchi, V., 1990. A history of flexible specialisation and industrial districts in EmiliaRomanga, in Pyke, F., Becattini, G. & Sengenberger, W., (eds.) 1990. Industrial districts and
inter-firm co-operaiton in Italy, International Institute for Labour Studies, Geneva, p. 10-19.
-
Casavola, P., Pellegrini, G. & Romagnano, E., 2000. Imprese e mercato del lavoro nei distretti
industriali, in Signorini, L.F., (eds.) 2000. Lo sviluppo locale Un’indagine della Banca d’Italia
sui distretti industriali. p. 51-65.
-
Centro Studi Confindustria Marche, 2008. Indagine congiunturale trimestrale,
commissioned by Banca Marche and Confindustria Marche, Ancona
-
Christaller, W., 1933. Die zentralen Orte in Süddeutschland. Eine ökonomisch-geographische
Untersuchung über die Gesetzmässigkeit der Verbreitung und Entwicklung der Siedlungen mit
städtischen Funktionen.
-
Coffey, A. & Atikinson, P., 1996. Making sense of qualitative data: Complementary research
strategies. Sage Publications, Thousand Oaks.
-
Collins, H., 2001. Tacit knowledge, Trust and the Q of Sapphire. Social studies of Sciences, 31
(1) p. 71-85.
-
Cohen, W. & Levinthal D., 1990. Absorptive Capacity: A new perspective on learning and
innovation. Administrative Science Quarterly, 35 (1) p. 128-152.
-
Conti, G., Cuccuelli, M. & Paradisi, M., 2007. Internationalizzazione e strategie delle imprese
nei settori tradizionali, L’Industria, 28 (1) p. 117-157.
-
Cucculelli, M., & Micucci, G., 2008. Family succession and firm performance: Evidence from
Italian family firms, Banca d’Italia, Rome.
-
D’Arpizio, C., 2008. Opportunità internazionali per il prodotto calzaturiero italiano: Scenari
odierni e prospettive future, Presentation commissioned by ANCI, Bain & Company, Lucca
71
-
De Blasio, G., & Di Addario, S., 2005. Do workers benefit from industrial agglomeration?
Journal of regional science 45 (4) p. 797-827.
-
Dei Ottati, G.,1987. Il mercato comunitario, in G. Becattini, Mercato e forze locali: il distretto
industrial. Bologna: Il Mulino p. 117-42.
-
Dei Ottati G., 1994. Cooperation and competition in the industrial district as an organization
model. European Planning Studies, 2(4).
-
Dei Ottati, G., 2002. Social concertation and local development: the case of industrial
districts. European Planning Studies, 10(4) p. 449-466.
-
Dhaliwal, D., 1986. Measurement of financial leverage in the presence of unfunded pension
obligations. The Accounting Review, 61(4) p. 651 – 661.
-
Dijk van, M., 1995. Flexible specialization, the new competition and industrial districts.
Small Business Economics, 7 p. 15-27.
-
Fabiani, S., Pellegrini, G., Romagnano, E. & Signorini, L.F., 2000. L’efficienza delle imprese
nei distretti industriali italiani, in Signorini, L.F., (eds.) 2000. Lo sviluppo locale Un’indagine
della Banca d’Italia sui distretti industriali.
-
Faulkner D., 2002. Critical perspectives on business and management, Routledge, London.
-
Farabullini, F. & Gobbi, G., 2000. Le banche nei sistemi locali di produzione, in Signorini,
L.F., (eds.) 2000. Lo sviluppo locale Un’indagine della Banca d’Italia sui distretti industriali.
-
Field, A., 2005. Dicovering statistics using SPSS, second edition, Sage publications, London.
-
Finaldi Russo, P. & Rossi, P., 2000. Costo e diponibilità del credito per le imprese nei
distretti industriali, in Signorini, L.F., (eds.) 2000. Lo sviluppo locale Un’indagine della Banca
d’Italia sui distretti industriali.
-
Finaldi Russo P. & Rossi, P., 2001. Credit constraints In Italian industrial districts. Applied
economics, 33, p. 1469-1477.
-
Gola, C. & Mori, A., 2000. Concentrazione spaziale della produzione e specializzazione
internazionale dell’industria italiana, in Signorini, L.F., (eds.) 2000. Lo sviluppo locale
Un’indagine della Banca d’Italia sui distretti industriali. p. 67-100.
-
Goodman, E., Bamford, J., Saynor, P., 1989. Small firms and industrial districts in Italy,
Routledge, London.
72
-
Gujarati, D., 2003. Basic econometrics, McGraw-Hill, New York.
-
Gray, S. & Needles, B., 1999. Financial accounting: a global approach, Houghton Mifflin,
Boston.
-
Grinblatt, M. & Titman, S., 2002. Financial markets and corporate strategy, McGraw-Hill,
New York.
-
Grove, A., 1996. Only the paranoid survive, Doubleday, New York.
-
Hadjimichalis, C., 2006. The end of third Italy as we knew it? Antipode, 38 (1) p. 82-106.
-
Harrison B., 1991. Industrial districts: Old wine in new bottles? Regional studies, 26 (5) p.
469-483.
-
Hendrikse, G., 2003. Economics and Management of Organizations. McGraw Hill, London.
-
ISTAT, 2001. I Distretti Industriali, in 8° Censimento generale dell’industria e dei servizi.
-
Jacobs, J., 1969 The economy of cities, Random House, New York.
-
Karlsson, C., Flensburg, P. & Horte S.A., 2004. Introduction of Knowledge spillovers and
knowledge management, Cheltenham: Edward Elgar, p. 3-18.
-
Krugman, P., 1991. Geography and Trade, MIT Press, Cambridge.
-
Luzzolino, G., 2000. I distretti industriali nel censimento intermedio del 1996: dimensioni e
caratteristiche strutturali, in Signorini, L.F., (eds.) 2000. Lo sviluppo locale Un’indagine della
Banca d’Italia sui distretti industriali.
-
Maddigan, R., 1981. The measurement of vertical integration. The Review of Economics and
Statistics, 63(3) p.328 -335.
-
Markusen, A., 1996. Sticky places in slippery space: A typology of Industrial Districts.
Economic Geography. 72(3) p. 293-313.
-
Marshall, A., 1890. Principles of economics: an introductory volume, McMaster University,
Candada.
-
Marshall, A., 1920. Industry and Trade. Macmillan, London.
73
-
McCann, P., Urban scale economies: statics and dynamics. In Cappelo, R. & Nijkamp, P.
(eds.) 2004. Urban dynamics and growth p. 31-56.
-
Modigliani, F. & Miller, M., 1958. The cost of capital, corporation finance and the theory of
investment. American Economic Review, 48(3) p. 261-297.
-
Murat, M. & Paba, S., 2004. International migration, outsourcing, and Italian industrial
districts, Working paper Università di Modena e Reggio Emilia
-
Ommicioli, M., 2000. L’indagine sul campo: una presentazione, in Signorini, L.F., (eds.) 2000.
Lo sviluppo locale Un’indagine della Banca d’Italia sui distretti industriali.p.289-298.
-
Pagano, M., 2000. Banche e distretti industriali: una relazione speciale?, in Signorini, L.F.,
(eds.) 2000. Lo sviluppo locale Un’indagine della Banca d’Italia sui distretti industriali.p.155167.
-
Paniccia, I., 1999. The Performance of IDs. Some insights from the Italian case. Human
systems management,18 p. 141-159.
-
Paniccia, I. 2002. Industrial Districts: evolution and competitiveness in Italian firms. Edgar
Elgar publishing limited, UK.
-
Pietrobelli, C., 1998. The Socio-Economic Foundations Of Competitiveness:
An Econometric Analysis of Italian Industrial Districts. Industry and Innovation, 5(2) p. 139155.
-
Porter, M., 2008. The five competing forces that shape strategy. Harvard business review
p.1-18.
-
Putnam, R., 1993. Making democracy work: Civic traditions in modern Italy, Princeton, NJ:
Princeton University Press.
-
Pindyck, R. & Rubinfield, D., 2003. Econometric models and Econometric forecasts, McGrawHill, New York.
-
Piore, M. & Sabel, C., 1984. The second industrial divide: possibilities for prosperity, Basic
Books, New York.
-
Pyke, F., Becattini, G. & Sengenberger, W., 1990. Industrial districts and inter-firm cooperation in Italy, International Institute for Labour Studies, Geneva.
74
-
Rabellotti, R., 1995. Is There an “Industrial District Model”? Footwear Districts in Italy and
Mexico Compared. World Development, 23(1). p. 29-41.
-
Rabellotti, R., Carabelli, A. & Hirsch, G., 2009. Italian Industrial Districts on the Move:
Where Are They Going? European Planning Studies, 17(1) p. 19-41.
-
Rosenthal, S. & Strange, W., 2003. Geography, industrial organization, and agglomeration,
The Review of Economics and Statistics, 85(2), p. 377-393.
-
Schumpeter, J., 1942. Capitalism, socialism and democracy, Harper, New York.
-
Signorini L.F., 1994. The Price of Prato, or measuring the industrial district effect. Papers in
Regional Science, 73 p. 369-392.
-
Signorini, L.F., 2000. Lo sviluppo locale Un’indagine della Banca d’Italia sui distretti
industriali, Meridiana Libri, Roma
-
Sforzi, F., 1989. The geography of industrial districts in Italy, in Goodman, E., Bamford, J. &
Saynor, P., (eds.) 1989. Small firms and industrial districts in Italy p. 153-166.
-
Sforzi, F., 1990. The quantitative importance of Marshallian industrial districts in the Italian
economy in Industrial districts and inter-firm co-operation in Italy, International Institute for
Labour Studies, Geneva.
-
Sforzi, F. 2002. The industrial district and the “New” Italian Economic Geography. European
Planning Studies, p.1-13.
-
Solinas, G., 1982. Labour market segmentation and workers’ careers: the case of Italian
knitwear industry, Cambridge Journal of Economics, 4.
-
Varaldo, R., 1988. Il sistema delle imprese calzaturiere, , Giappichelli, Torino.
-
Von Thünen, J., 1826. The isolated state.
-
Weber, A., 1909. Über den Standort der Industrie.
-
Williamson, O., 1975. Market and hierarchies, Free Press, New York.
-
Williamson, O., Wachter, M., & Harris, J., 1975. Understanding the employment relation: the
analysis of idiosyncratic exchange. Bell Journal of Economics, VI p. 250-278.
75
-
Williamson, O. 1985. The economic institutions of capitalism: firms, markets, relational
contracting, Free Press, New York.
-
Williamson, O., 2002. The theory of the firm as governance structure: Form choice to
contract, The Journal of Economic Perspectives 16(3) p. 171-195
-
Wippern, R., 1966. Financial structure and the value of the firm. The Journal of Finance 21(4)
p.615 – 633.
-
Wolfe A., & Gertler S., 2004. Cluster from the Inside and Out: Local Dynamics and Global
Linkages. Urban Studies, 41, p. 1071-1093.
-
Wooldridge, J., 2003. Introductory Econometrics: a modern approach, Thomson SouthWestern, United States.
-
Yin, K., 1994. Case Study Research: Design and Methods. Second edition, Applied Social
Research Methods Series, Volume 5, Sage publications.
8.2 Websites
- www.ancionline.com
- www.cna.it/eng
- www.confindustria.it
- www.confartigianato.it
- http://www.bvdep.com/en/Amadeus.html
Data extracted on November 19th 2008, Update number 181
76
9. Appendices
9.1 Questionnaire footwear companies
Introducing questions
1. What kind of activities does your company run?
2. Explain in what kind of production chain your company operates and the particular
role of your company within this production chain?
3. What is your own function within the company?
4. How old is the company?
5. Who founded the company?
(a) Is he/she still active in the company?
(b) If yes, in what kind of role?
Labor pool
6. How many workers does your company currently employ?
(a) Are those fulltime equivalent positions?
(b) Do you make use of temporary workers?
(c) What kind of contracts do your employees have (duration)?
7. How did you find your employees?
8. What is average time that it takes to fulfill an open vacancy?
9. How far do your employees live from the company?
Absorptive capacity
10. What is the average education level of your employees?
(distinguish for production and management roles)
11. What main experience / qualification do you look for in new employees?
(distinguish for production and management roles)
Family ownership
12. Does the company employ any family members?
13. What are their positions?
77
14. Are they registered as official employees or do they rather incidentally support your
business?
15. Do family members occasionally exceed regular working hours?
Boundaries of the firm
1. Suppliers
1. Did you use the same or different suppliers before 2007?
2. What is the average duration of your relationships with suppliers?
3. Did this change because of the economic crisis?
4. Did you ever position employees of yours at any of your suppliers?
(a) If yes, what was the reason for this?
(b) If yes, were you not afraid for leaking of knowledge?
5. What kind of contractual relationships do you have with your suppliers?
(a) If long term, did it ever happen that the suppliers broke the contract?
(b) How did you resolve this?
6. What is the geographical location of your main suppliers?
7. Do you have a different relation with geographically proximate suppliers compared to
more remotely located suppliers?
2. Customers
8. What kind of business are your customers in?
9. Did you have the same or a varying group of customers before 2007?
10. What is the geographical location of your main customers?
11. Do you have a different relation with geographically proximate customers compared
to more remotely located customers?
12. What kind of contractual relationships do you have with your customers?
(a) Did it ever happen that customers broke the contract?
(b) How did you resolve this?
3. Competition
13. What is the geographical location of your competitors?
14. How would you describe your relation with your competitors?
15. Do you have a different relation with geographically proximate competitors compared
to more remotely located competitors?
16. Did you ever cooperate with your competitors?
(a) Did you ever share materials with competitors?
(b) Did you ever share information and/or technological knowledge with any
competitors?
Institutions and policy
17. Which trade associations are you affiliated to?
78
18. What do you expect from these associations?
19. How do you regard the effectiveness of those associations?
20. What kind of services provided by policy institutions / trade associations would be
beneficial for your business?
21. How could institutions / trade associations improve to offer a beneficial service for
your business?
Financial intermediaries
22. Where do you obtain money from for long-term investments (for example in new
machinery)?
1 Family finance
(a)If family, why?
(b)If family, which family members (actively working in the company or not)?
2 Bank finance
(a) If bank, is it a local bank?
(b) Why did you choose this particular bank?
(c) If local bank, what kind of relationship do you have with the bank?
(d) How often do your interact with your contact person at the bank?
(e) Are you familiar with Confidi?
Location specific questions
23. Compared to a similar company in (outside) an agglomeration of footwear producers,
which specific advantages and disadvantages do you experience?
79
9.2 Overview tables and figures
Number
Title
2. Research subject
Figure 2.1
Main footwear producing agglomerations in Italy
Figure 2.2
Actors involved in the Fermano-Maceratese footwear district
3. Theoretical framework
Table 3.1
Explaining factors and proxies
Table 3.2
Features of Marshallian and Italianate industrial districts
Table 3.3
Four IIDs typologies
Figure 3.1
Porter’s five competitive forces framework
Table 3.4
Becattini’s distinctive IIDs features corresponding to Porter’s five forces
Table 3.5
Company performance indicators and their characteristics
Figure 3.2
Features related to company performance
Figure 3.3
Transaction governance
Table 3.7
Differences between flexible specialization and Fordist production
Figure 3.4
Degree of specialization determined by value added share
4. Methodology
Table 4.1
Variables overview
Figure 4.1
Dataset selection
Table 4.2
Assumptions ordinary least squares estimation
Table 4.3
Expert interviews
Table 4.4
Corporate interviews
Table 4.5
Operationalization of propositions
5. Results
Table 5.1
Table 5.2
Table 5.3
Table 5.4
Table 5.5
Table 5.6
Table 5.7
Table 5.8
Table 5.9
Descriptive statistics
T-test dependent variables
T-test explanatory variables
Correlation matrix
Multiple regression analysis
Multiple regression analysis (including interaction terms)
Quantitative results overview
Qualitative results overview
Results overview
80