Incubation matters. Measuring the effect
of business incubators on the innovation
performance of start-ups
Silvia Rita Sedita, Roberta Apa, Thomas Bassetti, Roberto Grandinetti
How to cite this paper:
Sedita, S. R., Apa, R., Bassetti, T., & Grandinetti, R. (2019). Incubation matters: Measuring the
effect of business incubators on the innovation performance of start-ups. R&D Management, 49(4),
This article explores the role of business incubators on the innovation performance of start-ups; in addition, we also
investigate how the incubation effect moderates other important factors driving their innovation performance. The
empirical evidence comes from a sample of firms located in Northern Italy belonging to the manufacturing (mechanical
engineering firms - MEF) and service sectors (knowledge-intensive business services—KIBS). The results suggest that
the incubation effect is very important in shaping the innovation performance of new ventures (measured as a percentage
of sales of new-to-market innovations). Moreover, it positively moderates the impact of a) the internal technical
capabilities and b) the adoption of a limited portfolio of collaborations for innovation.
This study explores the role of business incubators (BIs) in the determination of the innovation performance of start-
ups. BIs offer incubatees several facilities, from office space and capital to management support and knowledge (Allen
and Rahman, 1985; Sherman, 1999; Tamásy, 2007), qualifying potentially as a strong instrument to promote innovation
and entrepreneurship (Allen and McCluskey, 1990; Aerts et al., 2007), by reducing the “liability of newness"
Literature on BIs includes theoretically driven contributions, which provide BIs taxonomies (Hansen et al. 2000;
Grimaldi and Grandi, 2005; von Zedtwitz and Grimaldi, 2006), and qualitative research on the functioning of a BI,
which focuses on a) the services provided (Rice, 2002; McAdam and Marlow, 2007; Scillitoe and Chakrabarti, 2010); b)
the relationships among incubatees, and between incubatees and the incubator management (Bøllingtoft and Ulhøi,
2005; Schwartz and Hornych, 2010; Bøllingtoft, 2012; Ebbers, 2014; Rubin et al., 2015). Existing quantitative research
mainly debated the idea that start-ups located in a BI have a higher survival rate (Sherman, 1999; Ferguson and
Olofsson, 2004; Schwartz, 2013) and sales growth (Löfsten and Lindelöf, 2001; Colombo and Delmastro, 2002) as
compared to similar start-ups not located in a BI. Although the creation of BIs has been promoted by regional policies in
order to foster innovative entrepreneurship, there are few contributions that focused on the innovation performance of
incubatees (Colombo and Delmastro, 2002; Sullivan and Marvel, 2011). This work contributes to building awareness of
the existence, intensity and direction of the BI effect on the innovation performance of start-ups, providing original
Moreover, it looks at the innovation process under the lenses of the open innovation framework (Chesbrough, 2003),
which underlines the importance of the external to the firm network for innovation, when coupled with internal
knowledge investments and absorptive capacity. Accordingly, we also analyse if the engagement in an incubation
programme positively moderates the impact of internal capabilities and external collaborations on the innovation
performance of start-ups. Our empirical analysis is based on a sample of start-ups located in Northern Italy, operating
either in the manufacturing (mechanical engineering firms, MEF) or the service sectors (knowledge-intensive business
The originality of our work resides on: a) the adoption of a quantitative approach to the measurement of the impact of
BIs on the innovation performance of start-ups; b) the evaluation of how BIs moderate the impact of start-ups’ internal
capabilities on their innovation performance; c) the evaluation of how BIs moderate the impact of start-ups’ external
collaborations on their innovation performance; d) the analysis of start-ups belonging to both manufacturing and service
industries; and e) the analysis of a market-based measure of innovation performance – that is the commercialisation
The paper proceeds as follows. Section 2 illustrates theoretical background and hypotheses, Section 3 presents
methodological issues, Section 4 shows and discusses results, and Section 5 provides conclusions.
1. Theoretical background and hypotheses
The literature on BIs grew alongside two main streams of research: one focused on BI’s and the other on incubatees’
BIs’ performance is defined as the extent to which incubator outcomes correspond to incubator goals (Bergek and
Normann, 2008), and measured by several quality and efficiency measures (Hannon and Chaplin, 2003), mainly linked
to the evaluation of the activities conducted by the incubator organisation (Allen and McCluskey, 1990; Mian, 1997;
Löfsten and Lindelöf, 2001). Existing research mostly relies on qualitative evidence on the knowledge flows originated
and passed through the incubator, adopting a relational approach (Rice, 2002; Peters et al., 2004; Bøllingtoft and Ulhøi,
2005; Schwartz and Hornych, 2010; Scillitoe and Chakrabarti, 2010; Bøllingtoft, 2012; Ebbers, 2014). Literature also
provided analysis of the typologies of services and infrastructures that are more likely to improve incubatees’ daily
business (Tōtterman and Sten, 2005; Soetanto and Jack, 2013).
Incubatees’ performance is generally estimated by survival rate (Allen and McCluskey, 1990; Mian, 1997; Peña,
2004; Aerts et al., 2007; McAdam and Marlow, 2007; Bruneel et al., 2012); sales and employment growth rate
(Colombo and Delmastro, 2002; Peña, 2004; Sung, 2007; Soetanto and Jack, 2013); profit growth rate (Mian, 1997;
Chen, 2009); export growth rate (Mian, 1997); satisfaction with the return on asset (Chen, 2009).
Existing literature has not considered innovation performance as a measure of incubatee’s performance except for
Colombo and Delmastro (2002) and Sullivan and Marvel (2011). In particular, Sullivan and Marvel (2011) investigated
the relationship between knowledge acquisition and the innovativeness of the initial product/service of early-stage
technology start-up operating in university-affiliated incubators, not mentioning the role of BIs in fostering innovation
performance. Differently, Colombo and Delmastro (2002), by considering a technology BI (Phillips, 2002) – reported
that incubatees show a superior innovation performance compared to off-incubator firms. The analysis of technology BIs
is interesting and promising; nevertheless, the concept of innovation transcends that of technology-intensive activities.
Many innovations are not technology driven, but related, for instance, to business models changes. The analysis of the
impact of BIs on the innovation performance has been so far limited to high-tech start-ups; it is, therefore, important to
provide evidence, as we do, for other categories of start-ups. Building on the insights provided by Colombo and
Delmastro (2002), we formulate our baseline hypothesis (H1).
H1. The engagement in an incubation programme positively affects the innovation performance of start-ups.
The importance of open innovation as input for new business models is widely recognised (Chesbrough, 2003). For
succeeding in creating and selling innovative products and services is necessary to create and deploy successful
collaborations with competitors, consultants, suppliers, clients, universities and other research organizations (Laursen
and Salter, 2006; Baba et al., 2009). A largely established view in the management and economics literature underlines
the complementary nature of internal capabilities and external collaborations (Arora and Gambardella, 1994; Powell et
al., 1996; Cassiman and Veugelers, 2006). Through recombinant capabilities, firms acquire and use complementary
knowledge to generate technological innovation (Galunic and Rodan, 1998; Yayavaram and Ahuja, 2008). In order to
work successfully, collaborative innovation models require firms building up networking and absorptive capabilities
(Cohen and Levinthal, 1990), which derive from previous experience and repeated interactions with specific members of
the value chain (clients and/or suppliers), or established routines of exploration activities that transcend problem solving
and daily business. The liability of newness affects also innovation activities (Stinchombe, 1965), and incubators can
overcome the limits of start-ups influencing the way they develop internal capabilities and networks. The innovation
performance of firms relies, in fact, on the availability of proper capabilities to enhance innovative efforts (Porter, 1980;
Lee et al., 2001). Following Lee et al. (2001), we classify capabilities into business and technical. The interaction among
incubatees and between incubatees and the incubation management allows for a better use of internal capabilities,
avoiding strategic failures. This positive feedback process between incubatees’ capabilities and BI’s services emerged
clearly from previous empirical works (Rice, 2002; Scillitoe and Chakrabarti, 2010; Monsson and Jörgensen, 2016; Apa
et al., 2017). Despite the importance of this intuition, existing research did not compare the performance of incubatees
with off-incubator start-ups with similar capabilities. On the basis of this literature, we first test the direct effect of
capabilities on innovative performance (H2.1), and then the moderating effect of being engaged in an incubation
H2.1: Business and technical capabilities positively affect the innovation performance of start-ups.
H2.2: The engagement in an incubation programme moderates positively the impact of business and technical
capabilities on the innovation performance of start-ups.
Start-ups are unable to invest heavily in internal R&D activity, due to lack of financial resources, and often miss the
point when trying to commercialise new products, due to lack of marketing capabilities. It follows that they cannot count
only on their internal knowledge; on the contrary, they have to build fruitful relationships with other organisations,
which give them a chance to multiply their learning opportunities (Sedita and Apa, 2016). BI helps incubatees to
develop their networks fuelling connections among incubatees and stimulating them to build useful external-to-the-BI
relationships in two ways.
First, BI provides incubatees with resources, capabilities, knowledge, learning and social capital that are useful to
manage efficiently their network. BIs facilitate the management of different types of relationships (Phelps et al. 2012),
increasing the probability to transform a friendship into a profitable business relationship. Second, BI provides the
adequate support in scouting and selection processes of possible partners. After the seminal contribution of Hansen et al.
(2000), many scholars studying BIs shared the idea that the most efficient incubation model is that of the networked
business incubator, which works as a relationships enabler. This conclusion has been reached through a variety of
theoretical and qualitative research analyses (mainly based on case studies) (Bøllingtoft and Ulhøi, 2005; Tōtterman and
Sten, 2005; McAdam and McAdam, 2006; Bøllingtoft, 2012; Sá and Lee, 2012; Soetanto and Jack, 2013; Rubin et al.,
2015; Apa et al., 2017). BI, having a higher knowledge of the market and the environment in which the start-ups are
involved, is able to provide incubatees with a list of possible partners that could fit their needs at the best, increasing the
quality of the overall incubatees’ network. Following Laursen and Salter (2006), we assume that the breadth of external
collaborations is positively associated with the innovation performance. Moreover, a large number of high-quality
relationships increase recombination opportunities of different knowledge and competencies that lead to better
innovation performance. We first test the direct effect of networks on innovation performance (H3.1), and then the
moderating effect of being engaged in an incubation programme (H3.2).
H3.1: Collaborations positively affect the innovation performance of start-ups.
H3.2: The engagement in an incubation programme positively moderates the impact of collaborations on the
innovation performance of start-ups.
The data for the analysis come from an original survey conducted between February and June 2013. Our unit of analysis
consists of all of the shared capital companies registered within the firms’ register of the Italian Chambers of Commerce
that comply with all the following characteristics: 1) born between 2005 and 2009; 2) still active in 2012; 3) located in
Northern Italy, within any of the following regions: Emilia-Romagna, Friuli-Venezia Giulia, Trentino Alto-Adige and
Veneto (North-East), or Liguria, Lombardia, Piemonte and Valle D’Aosta (North-West); 4) either being KIBS or MEF.
The choice of these two industries is motivated by their crucial contribution to the competitiveness of Italian economy
and the fact they represent two mid-tech industries, rather understudied in previous research works in this field. Such
sectors are also interesting because of their relevance for other industries, considering that their output, being a service
or a product, supports the innovativeness and performance of clients and intermediaries. Table 1 reports the specific
ATECO 2002 (the Italian version of the NACE rev.1) codes. Out of this population of firms, we drew our sample
adopting a stratified sampling procedure by industry specialisation and geographical location (regional level).
Data were collected using a mixed mode design. A specialised survey company conducted the interviews with the
assistance of CAWI (Computer Assisted Web Interviewing) and CATI (Computer Assisted Telephone Interview)
procedures, targeting the entrepreneurs. As highlighted by De Leeuw (2005), mixed modes constitute an affordable way
to compensate for the weakness of each individual mode. Indeed, mixed designs let respondents choose their favourite
interview method, minimising non-response and non-response bias. In order to correct for CAWI self-selection effects
and ensure the representativeness of the sample, we first used the CAWI method collecting 39% of total interviews and
then the CATI method which accounted for the remaining 61% of interviews. From the 2,341 firms initially contacted,
we collected 409 interviews. However, because some responses were missing, the number of observations dropped to
243 (128 are KIBS and 115 are MEF). To test for non-response bias within the sample, we performed multivariate tests
on means, correlations and covariances. According to the results reported in Appendix A, we cannot reject the
hypothesis of equality between the response and the non-response group.
INSERT TABLE 1 ABOUT HERE
3.2.1 Dependent variable
Start-ups are often evaluated by means of survival (Aerts et al., 2007), growth (Soetanto and Jack, 2016) or innovation
performance indicators, such as patent activity or number of patents (Colombo and Delmastro, 2002; Ferguson and
Olofsson, 2004). Since we also consider start-ups operating in the service sectors, these traditional measures of
innovative effort may not be adequate. For this reason, we measure the innovation performance through the fraction of
the turnover related to products new to the market (INNO_PERF). This measure is both cross-sectorial (allowing for
comparative studies between manufacturing and service firms) and closer to the real value that a company receives from
its innovative effort (Teece, 1986; Schneider and Veugelers, 2010). It is in fact well acknowledged that being innovation
leaders is not necessarily connected to being able to make profits from innovation (Teece, 1986). A more consistent
measure of innovation performance is required in order to capture the ability of the firm not only to innovate, but also to
make profits from innovation.
Since respondents tend to round to the nearest 5%, the fraction of the firm’s turnover relating to products new to the
market has too many categories with a small number of observations per category. To avoid potential problems with the
estimates of standard errors, we classify the dependent variable into five classes corresponding to the quintiles of the
distribution of the innovation performance. This allows us to obtain results that are robust to rounding and that can be
3.2.2 Main independent variables
In order to assess if the engagement in an incubation programme affects start-up innovation performance we introduced
the variable INCUB, which assumes a value of 1 if the firm passed through an incubator and 0 otherwise.
In order to assess the impact of firm-specific factors on the innovation performance of firms, we evaluated the set of
capabilities owned by the firm after three years of activity. As from the questionnaire items, we distinguished between
four types of capabilities: technological (TECH), marketing (MKTG), management (MAN), and ICT. Respondents were
asked to indicate the degree of their capabilities on a five-point scale ranging from 1 (no capabilities) to 5 (extremely
high capabilities). Since firms having high managerial capabilities are also likely to have high marketing capabilities
and, similarly, technical capabilities might be correlated with ICT capabilities (see Table A2 in Appendix A), we used a
principal component analysis (PCA) to obtain two composite indices capturing business (BUSINESS) and technical
capabilities (TECHNICAL) owned by the firm. By using these two indices instead of the original variables, we mitigate
the possibility of measurement errors coming from self-reported data.
Stemming from the seminal contribution of Laursen and Salter (2006), several empirical works quantitatively measured
the open innovation strategy of firms by describing the breadth of the inbound information sources and/or the innovation
collaborations (e.g., van der Meer, 2007; Frenz and Ietto-Gillies, 2009; Leiponen and Helfat, 2010; Sofka and Grimpe,
2010; Love et al., 2011). Accordingly, we introduce a variable reflecting the variety of collaboration channels through
which a start-up is seeking to accumulate external resources. The variable is termed COLL and is constructed as a
combination of five types of collaborations for innovation (suppliers, clients, competitors, consultants, universities and
other public research organisations). As a starting point, each of the five types is coded as a binary variable, 0 being no
use and 1 being use of the given partner. Subsequently, the five types of partnerships were simply added up so that a
firm gets a 0 when no collaborations are in place, while a firm gets a value of 5 when all types of collaborations are
used. In other words, it is assumed that firms that use a higher variety of partnerships are more ‘open’, with respect to
external collaboration strategies, than firms that do not. In order to take into account the discrete nature of the variable as
well as possible threshold effects and nonlinearities, we treat COLL as a factor variable. Indeed, treating COLL as a
continuous variable would imply the aprioristic assumption that passing from no collaboration to 1 collaboration has the
same impact on the innovative capacity as passing from 3 to 4 or from 4 to 5.
3.2.3 Control variables
Intellectual property (IP) protections
Protection mechanisms aim at guaranteeing the appropriation of returns deriving from a firm’s innovative activity
(Hertzfeld et al., 2006). In general, the effectiveness of these mechanisms depends on the type of innovation, the
industry and the firm size. Manzini and Lazzarotti (2015) recently studied which kinds of IP protection mechanisms
should be used in different phases of new product development. On the basis of these contributions, and observing the
importance of appropriability when adopting an open innovation strategy (Laursen and Salter, 2014), we considered the
number of protections used by the firm by introducing the variable PROT. This variable is constructed from responses to
a specific question and is a combination of four types of protections (trademarks, patents, designs and trade secrets). As
a starting point, each of the four types was coded as a binary variable, 0 being no use and 1 being use of the given
protection. Subsequently, the four types of protections were added up so that a firm gets a 0 when no protections are
used and a value of 4 when all types of protections are used.
Traditionally, one of the key strategies to secure technological potential, and, therefore, innovation and economic
growth, is investment in R&D (Trajtenberg, 1990). R&D investment increases the possibility of achieving a higher
standard of technology in firms and regions, which would allow them to introduce new and superior products and/or
processes, resulting in higher levels of income and growth (Wakelin, 2001; Bilbao-Osorio and Rodríguez-Pose, 2004).
We assess the impact of R&D expenditure on innovation performance by considering the (self-reported) percentage of
turnover invested in R&D, measured in the first year of the start-up’s life.
In Italy, the industrialisation process has shown different regional patterns (Quatraro, 2009). Since diverse geographical
areas provide different institutional settings that may influence the innovation performance of firms, we introduce the
control variable GEO, which takes a value of 1 if the firm is located in the North-Western part of Italy, and 0 if the firm
is located in the North-Eastern one.
In order to control for different levels of innovation performance across sectors, we introduced a dummy variable
INDUSTRY, which takes value of 1 if the firm is a KIBS and a 0 if the firm is a MEF.
Firm size and age
The existing literature suggests that both the age and the size of the firm are crucial predictors of its survival capacity
(see, e.g., Ferguson and Olofsson, 2004). At the same time, the empirical evidence on the relationship between firm size
and innovation performance is extremely mixed (see, e.g., Kamien and Schwartz, 1982; Kleinknecht and Reijnen, 1991;
Vaona and Pianta, 2008), while the probability of innovation tends to be higher in young firms (Huergo and Jaumandreu,
2004). For these reasons, we took into account both firm size and age. SIZE corresponds to the number of active
founders and employees in 2012. AGE is derived from the date of foundation of the company, and (being calculated in
2012) spans from 3 to 7, meaning that the oldest firms were founded in 2005 and the youngest in 2009.
Type of new venture
New ventures are usually classified as either start-ups, or academic/corporate spin-offs. The nature of the new venture
affects the innovation strategy (Grimaldi and Grandi, 2005), therefore we created the variable TYPE, which takes a
value of 1 if the company is a spin-off and 0 otherwise.
Internationalisation may help firms accumulate important inputs for innovation (Kobrin, 1991). According to Kotabe
(1990), export-oriented firms have access to a wider range of resources that improve their innovative capacity. For
instance, exporters may take advantage of strategic alliances with foreign suppliers, institutions and competitors (Hitt et
al., 1997). In addition, firms competing in international markets are also exposed to international technological spillovers
(Liu and Buck, 2007). EXPORT is a dummy variable, which takes a value of 1 if the company reports some foreign
sales and 0 otherwise.
3.3 The model
We first test our hypothesis H1, by considering only the direct relationship between INCUB and the dependent variable
(INNO_PERF). The empirical analysis is based on two standard regression models (Table 4). Since our dependent
variable is far from being normally distributed, we estimate an ordered Logit model in which the thresholds are known
(we consider five equally numerous classes). We then test the remaining hypotheses by including in the model
estimation a set of interaction terms (Table 5). Because INCUB is a dichotomous variable, the interpretation of
interaction effects is straightforward. In particular, in Table 5, we tested hypothesis H2.2 by including in column 1 an
interaction term between INCUB and two variables indicating whether the firm owns BUSINESS and/or TECHNICAL
capabilities. Similarly, we test H3.2 by introducing an interaction term between INCUB and a factor variable indicating
the number of collaboration channels. Table 2 lists the variable descriptions and Table 3 shows the descriptive statistics.
INSERT TABLE 2 ABOUT HERE
INSERT TABLE 3 ABOUT HERE
4. Results and discussion
Table 4 reports the ordered Logit estimates for the percentage of revenues coming from new products in 2012. In
column 1, we estimate the model without independent variables. Only two control variables are significantly correlated
with the firm’s innovation performance: the firm’s age and its initial R&D expenditure. In particular, there is a positive
relationship between the initial R&D expenditure and the dependent variable, and a negative relationship between the
firm’s age and the innovation performance. However, this second result is due to the small number of firms that were
born in 2009.
In column 2, we test hypothesis H1. According to our results, those firms that engaged in an incubation programme
exhibit a higher innovation performance. In other words, incubators facilitate the returns from the innovation process,
thus confirming H1. This result is crucial and works as baseline for developing the further hypotheses testing exercise.
As in Colombo and Delmastro (2002), the incubation effect on innovation performance of start-ups exists and the
magnitude is particularly relevant. Notice that the introduction of the incubation effect sensitively increases the pseudo-
R2, that is, the inclusion of the incubation effect improves the model fit.
In column 3, together with hypothesis H1, we also test hypotheses H2.1 and H3.1. Regarding internal capabilities,
business and technical capabilities per se are not significantly associated with innovation performance. Thus H2.1 is not
confirmed. This is surprising, since a lack of internal capabilities is often considered as a major issue in discriminating
innovative new ventures (e.g., Park, 2005; De Winne and Sels, 2010). Nevertheless, start-ups are often unable to
evaluate properly their capacities, incurring in the risk of over-estimating their internal resources. Since we collected the
information about the level of capabilities through self-reported data, it reflects the perception of the entrepreneur, which
is obviously affected by his/her reference system. Being a start-up means not having experience in the competitive
environment, which allows to benchmark with other companies and to receive feedback from peers/experts. This leads
start-ups not to evaluate correctly their capabilities. In a similar vein, new ventures may not be able to recognize the
complementary capabilities needed in order to transform their resources in successful new products and services. In our
analysis, the two types of capabilities (business and technical) derive from a PCA, therefore they are orthogonal, and not
correlated. There is not a clear relation between the level of business and of technical capabilities. We might infer that
start-ups should increase their business/technical capabilities to complement their initial set of capabilities. In other
words, start-ups seem to need an external support for more carefully evaluate their strengths and weaknesses and for
developing further those capabilities that allow them to take advantage of the opportunities in the market and avoiding
the risks related to the threats.
Reading Table 4, collaboration breadth seems not to affect the firms’ innovation performance. A possible explanation
for this lack of evidence could be the relatively young age of firms entering our sample. Start-ups may be too young to
have a solid and effective network of collaborations with clients and suppliers (also discussed by Stinchcombe, 1965,
referring to the liability of newness phenomenon). Indeed, collaborations tend to take time to be developed. Moreover,
the magnitude of the initial R&D intensity positively affects the innovation performance. At a first glance, it seems that
the open innovation argument for start-ups is not working and that the importance of the internal R&D investments
overcomes the networking abilities. Thus H3.1 is not confirmed. This result paves the way to a more careful debate on
the role of open innovation for start-ups, since the phenomenon has been prevalently looked under the lenses of the large
corporation, which often “uses” start-ups for knowledge exploration through acquisition processes (Barkema and
Vermeulen, 1998), corporate incubators (Becker and Gassmann, 2006), crowdsourcing platforms and on-line
competitions (Lampel et al. 2012). Moreover, the analysis of innovation collaborations for start-ups are so far limited to
the potential given by the participation to on-line communities (West and Lakhani, 2008), especially in the case of the
software industry (Autio et al., 2013). Finally, the incubation effect remains visible and the only one affecting the
innovation performance of start-ups, being networking and internal capabilities not significant. According to the results
reported in Table 4, in order to have the same innovation performance guaranteed by the engagement in an incubation
programme, a firm must increase its initial R&D expenditure by 18%.
INSERT TABLE 4 ABOUT HERE
Results deriving by testing H2.1 and H3.1 deserve further attention in order to disentangle more clearly how eventually
being part of an incubation programme changes the way networking activities are conducted and internal capabilities are
combined and used. We expect to find a significant role of the BI in fuelling and selecting appropriate network partners
and exploiting properly internal capabilities. Therefore, in Table 5, we test hypotheses H2.2 and H3.2 thus verifying the
moderating effect of BIs on internal capabilities and on the collaboration breadth in shaping start-ups innovation
performance. With respect to column 3 of Table 4, in column 1 of Table 5, we interact the incubation dummy with two
variables capturing business (BUSINESS) and technical (TECHNICAL) capabilities. The coefficient of the interaction
term INCUB*TECHNICAL is positive and statistically significant at 5%. This means that high technical capabilities
have a positive impact on innovation performance only for those firms that passed through an incubator. Owning
technical capabilities per se is not sufficient to transform them into successful new-to-the market products. This is
consistent with an innovation approach that considers as highly relevant that the product design phase is parallel to a
market analysis phase, which provides the right feedback to increase the fitness between the innovative product and the
demand needs. This process is eased by the support of the management team of a BI, which overcomes the lack of
experience of a start-up. Being incubated allows start-ups to better assess their strengths and weaknesses, and this
contributes to explain the moderating effect we observe in Table 5. Again, the moderation effect showed in Table 5
demonstrates how the experience within an incubator supports firms in providing complementary resources. Therefore,
H2.2 is confirmed. As it has been also discussed by the literature on BI (Rice, 2002), the provision of support services is
useful for increasing the incubatees’ abilities in many directions. Our result qualifies as the first empirical evidence of
the moderating effect of BI on the capacity of start-ups to generate profits from innovation.
In column 2 of Table 5, we interact the incubation dummy with the number of types of collaborations reported by the
firm. In particular, by considering the ordinal nature of COLL, we find that incubators enhance the innovation
performance of those firms characterised by two or three types of collaborations. Given our specification, the coefficient
for the sole variable INCUB captures the average innovation performance of firms that have been incubated, but are not
engaged in any type of collaborations. Notice that, despite the number of these firms is too small to draw any reliable
conclusion, their performance does not significantly differ from that of firms engaged in all types of collaborations.
Results reported in column 2 of Table 5 suggest that the incubation effect found in Table 4 is mainly due to the fact that
firms passing through incubators tend to develop collaborations that boost their innovation performance. A possible
explanation for this result might be due to the quality of collaborations: BIs are able to orient the start-ups towards the
selection of the most valuable partners and to push start-ups’ products into the proper market, maximising the returns
from the innovation effort. Thus H3.2 is confirmed. Open innovation strategies are beneficial for start-up companies if
they pass through an incubation experience. This result is new and enriches the present understanding on how to sustain
the innovation performance of start-ups through the adoption of an open innovation strategy orientation.
INSERT TABLE 5 ABOUT HERE
Overall, an industry effect arises. In particular, KIBS show an innovation performance lower than that of MEF.
This work investigated the impact of being engaged in an incubation programme on the innovation performance of
start-ups, measured as the fraction of returns from products new to the market. Despite many theoretical and empirical
studies considered the engagement in an incubation programme as an effective tool for encouraging the birth and the
survival of technological innovative new ventures (Sherman, 1999; Ferguson and Olofsson, 2004; Schwartz, 2013), few
studies have measured the impact of BIs on start-ups’ innovation performance (Colombo and Delmastro, 2002; Sullivan
and Marvel, 2011). By using data from a survey of MEF and KIBS start-ups in Northern Italy, we demonstrated that: a)
consistently with Colombo and Delmastro (2002), firms that spent a period within a BI show a better innovation
performance compared to other start-ups; b) BIs play a moderating role by enhancing the effects of internal capabilities
on start-ups’ innovation performance, in particular, technical capabilities (technology and ICT); c) BIs positively
moderate the effect of collaboration breadth on innovation performance, when start-ups report an intermediary
collaboration breadth level. Therefore there is a peculiar behaviour of start-up companies conducive to superior
performance in terms of profits from innovation: they have to pass through a BI, which is able to recommend a better
evaluation and use of internal capabilities, and provide the adequate support in the scouting and selection process of
possible business and innovation partners. The literature on BIs is therefore enriched by our contribution, which not only
demonstrates that being incubated is beneficial for start-ups, but also that is fundamental in order to adopt successfully
an open innovation business model. Exploration and exploitation activities (Gupta et al., 2006), in fact, are empowered
by the services offered by a BI. This is a completely new and original result for the literature on BI and start-ups, which
never provided empirical evidence of the manner through which BI guarantees to start-ups to profit from innovation.
Given this interesting spectrum of results, we can draw some policy implications. The investment in incubation
structures oriented towards improving networking and internal capabilities is particularly useful for sustaining successful
innovative entrepreneurship. BIs configure themselves as an essential element for boosting the competitive advantage of
regions in a knowledge economy. Therefore, entering into the debate on the usefulness of BIs, we strongly recommend
policymakers to support the birth and the development of these organisations, while also carefully looking at the services
provided, so as not to generate or keep alive ghost structures unable to create economic value. Moreover, the empirical
evidence here provided suggest to prefer the networked BI (Hansen et al., 2000) as a successful model of incubation,
because it seems to be the most qualified in order to increase the networking abilities of incubatees (Eveleens et al.,
2017), thus preparing them to better adopt a successful open innovation business model. This is a key issue, especially in
the case of public BIs, which do not have a profit orientation, thus facing the risk of being inefficient. Further research
on the evaluation of BIs and the relative performance of public vs. private BIs, through appropriate measures, is
This research has some limitations. Firstly, although the literature shows that different types of BIs can sometimes
more or less support the innovation performances of start-ups (Hackett and Dilts, 2004; Grimaldi and Grandi, 2005;
Mian et al., 2016), the questionnaire used for the survey did not take into consideration the type of incubator in which
the firm was hosted. Secondly, since we have no information on pre-incubation capabilities and ties, we cannot say
whether BIs help start-ups accumulate internal and external resources or exploit existing ones. Future research should
investigate this important issue. Other limitations include country specificity and self-reported evaluation of some
variables used in the quantitative analysis. Nevertheless, the article sheds light on a phenomenon that has been
completely neglected in the literature, which is the market performance due to products new to the market of start-ups
that benefit from an incubation programme. It adds significantly to the literature on BIs, which is typically bounded to
the analysis of incubated firms, without comparing their performance with stand-alone firms. Moreover, it offers some
advice for policymakers and entrepreneurs seeking to sustain innovative entrepreneurship and profits from innovation.
We gratefully acknowledge valuable and stimulating comments from participants to the R&D Management
Conference 2015 in Pisa on a previous version of this work. We wish to thank the editor and the anonymous referees,
who provided precious comments that helped a lot improving the manuscript. The usual disclaimer applies. This work
was supported by the University of Padova (grant number CPDA142857/14 and CPDA115589/11).
Aerts, K., Matthyssens, P. and Vandenbempt, K. (2007) Critical role and screening practices of European business
incubators. Technovation, 27, 5, 254–267.
Allen, D.N. and McCluskey, R. (1990) Structure, policy services, and performance in the business incubator industry.
Entrepreneurship: Theory and Practice, 15, 2, 61–77.
Allen, D.N. and Rahman, S. (1985) Small business incubators: a positive environment for entrepreneurship. Journal of
Small Business Management, 23, 3, 12–23.
Apa, R., Grandinetti, R. and Sedita, S.R. (2017) The social and business dimensions of a networked business incubator:
the case of H-Farm. Journal of Small Business and Enterprise Development, 24, 2, 198–221.
Arora, A. and Gambardella, A. (1994) The changing technology of technological change: general and abstract knowledge
and the division of innovative labour. Research Policy, 23, 5, 523-532.
Arundel, A. and Kabla, I. (1998) What percentage of innovations are patented? Empirical estimates for European firms.
Research Policy, 27, 2, 127–141.
Autio, E., Dahlander, L. and Frederiksen, L. (2013). Information exposure, opportunity evaluation, and entrepreneurial
action: An investigation of an online user community. Academy of Management Journal, 56, 5, 1348-1371.
Baba, Y., Shichijo, N. and Sedita, S. R. (2009) How do collaborations with universities affect firms’ innovative
performance? The role of 'Pasteur scientists' in the advanced materials field. Research Policy, 38, 5, 756-764.
Barkema, H. G., & Vermeulen, F. (1998). International expansion through start-up or acquisition: A learning perspective.
Academy of Management Journal, 41(1), 7-26.
Bauer, T.K. and Sinning, M. (2008) An extension of the Blinder-Oaxaca decomposition to nonlinear models. Advances in
Statistical Analysis, 92, 2, 197–206.
Becker, B. and Gassmann, O. (2006). Corporate incubators: industrial R&D and what universities can learn from them.
The Journal of Technology Transfer, 31, 4, 469-483
Bergek, A. and Norrman, C. (2008) Incubator best practice: a framework. Technovation, 28, 1-2, 20-28.
Bilbao-Osorio, B. and Rodríguez-Pose, A. (2004). From R&D to innovation and economic growth in the EU. Growth
and Change, 35, 4, 434–455.
Bøllingtoft, A. (2012) The bottom-up business incubator: leverage to networking and cooperation practices in a self-
generated, entrepreneurial-enabled environment. Technovation, 32, 5, 304–315.
Bøllingtoft, A. and Ulhøi, J.P. (2005) The networked business incubator: leveraging entrepreneurial agency? Journal of
Business Venturing, 20, 2, 265–290.
Bruneel, J., Ratinho, T., Clarysse, B. and Groen, A. (2012) The evolution of business incubators: Comparing demand and
supply of business incubation services across different incubator generations. Technovation, 32, 2, 110–121.
Cassiman, B. and Veugelers, R. (2006) In search of complementarity in innovation strategy: Internal R&D and external
knowledge acquisition. Management Science, 52, 1, 68-82.
Cefis, E. O. and Marsili (2005) A matter of life and death: innovation and firm survival. Industrial and
Corporate Change, 14, 6, 1167–1192
Chen, C.-J. (2009) Technology commercialization, incubator and venture capital, and new venture performance. Journal
of Business Research, 62, 1, 93–103.
Chesbrough, H. (2003) Open Innovation: The New Imperative for Creating and Profiting from Technology. Boston:
Harvard Business School Press.
Cho, H.-C., Abe, S. (2013) Is two-tailed testing for directional research hypotheses tests legitimate? Journal of Business
Research, 66, 9, 1261–1266.
Cohen, W. M. and Levinthal, D. A. (1990) Absorptive capacity: A new perspective on learning and innovation.
Administrative Science Quarterly, 35, 1, 128–152.
Colombo, M.G. and Delmastro, M. (2002) How effective are technology incubators? Evidence from Italy. Research
Policy, 31, 7, 1103–1122.
De Leeuw, E.D. (2005) To mix or not to mix data collection modes in surveys. Journal of Official Statistics, 21, 2, 233–
De Winne, S. and Sels, L. (2010) Interrelationships between human capital, HRM and innovation in Belgian start-ups
aiming at an innovation strategy. The International Journal of Human Resource Management, 21, 11, 1863–1883.
Ebbers, J.J. (2014) Networking behavior and contracting relationships among entrepreneurs in business incubators.
Entrepreneurship: Theory and Practice, 38, 5, 1159–1181.
Eveleens, C.P., van Rijnsoever, F.J. and Niesten, E.M. (2017) How network-based incubation helps start-up
performance: a systematic review against the background of management theories. Journal of Technology Transfer,
42, 3, 676-713.
Ferguson R. and Olofsson, C. (2004) Science parks and the development of NTBFs: location, survival and growth.
Journal of Technology Transfer, 29, 1, 5–17.
Frenz, M. and Ietto-Gillies, G. (2009) The impact on innovation performance of different sources of knowledge: evidence
from the UK Community Innovation Survey. Research Policy, 38, 7, 1125–1135.
Galunic, D. C. and Rodan, S. (1998) Resource recombinations in the firm: knowledge structures and the potential for
Schumpeterian innovation. Strategic Management Journal, 19, 2, 1193–1201.
Grimaldi, R. and Grandi, A. (2005) Business incubators and new venture creation: an assessment of incubating models.
Technovation, 25, 2, 111–121.
Gupta, A. K., Smith, K. G. and Shalley, C. E. (2006). The interplay between exploration and exploitation. Academy of
Management Journal, 49, 4, 693-706.
Hackett, S.M. and Dilts, D.M. (2004) A systematic review of business incubation research. Journal of Technology
Transfer. 29, 1, 55–82.
Hannon P. and Chaplin P. (2003) Are incubators good for business? Understanding incubation practice the challenges for
policy’, Environment and Planning C: Government and Policy, 21, 6, 861–881.
Hansen, M.T., Chesbrough, H.W., Nohria, N. and Sull, D.N. (2000) Networked incubators: hothouses of the new
economy. Harvard Business Review, 78, 5, 74–84.
Hertzfeld, H.R., Link, A. and Vonortas, N.S. (2006) Intellectual property protection mechanisms in research partnerships.
Research Policy, 35, 6, 825–838.
Hitt, M., Hoskisson, R. and Kim, H. (1997) International diversification: effects on innovation and firm performance in
product-diversified firms. Academy of Management Journal, 40, 4, 767–798.
Huergo, E. and Jaumandreu, J. (2004) How does probability of innovation change with firm age? Small Business
Economics, 22, 3, 193–207.
Kleinknecht, A. and Reijnen, J.O.N. (1991) More evidence on the undercounting of small firm R&D, Research Policy,
20, 6, 579–587.
Kobrin, S. (1991) An empirical analysis of the determinants of global integration. Strategic Management Journal, 12, S1,
Lampel, J., Jha, P. P., and Bhalla, A. (2012). Test-driving the future: How design competitions are changing innovation.
The Academy of Management Perspectives, 26, 2, 71-85.
Laursen, K. and Salter, A. (2006) Open for innovation: the role of openness in explaining innovation performance among
UK manufacturing firms. Strategic Management Journal, 27, 2, 131–150.
Lee, C., Lee, K. and Pennings, J.M. (2001) Internal capabilities, external networks, and performance: a study on
technology-based ventures. Strategic Management Journal, 22, 6-7, 615–640.
Liu, X. and Buck, T. (2007) Innovation performance and channels for international technology spillovers: evidence from
Chinese high-tech industries. Research Policy, 36, 3, 355–366.
Löfsten, H. and Lindelöf, P. (2001) Science parks in Sweden: industrial renewal and development. R&D Management,
31, 3, 309–322.
Manzini, R. and Lazzarotti, V. (2015) Intellectual property protection mechanisms in collaborative new product
development. R&D Management, in press DOI: 10.1111/radm.12126.
McAdam, M. and Marlow, S. (2007) Building futures or stealing secrets? Entrepreneurial cooperation and conflict within
business incubators. International Small Business Journal, 25, 4, 361-382.
McAdam, M. and McAdam, R. (2006) The networked incubator: the role and operation of entrepreneurial networking
with the university science park incubator (USI). International Journal of Entrepreneurship and Innovation, 7, 2, 87–
Mian, S.A. (1997) Assessing and managing the university technology business incubator: An integrative framework’,
Journal of Business Venturing, 12, 4, 251–285.
Mian, S., Lamine, W. and Fayolle, A. (2016) Technology business incubation: an overview of the state of knowledge.
Technovation, 50-51, 1–12.
Monsson, C.K. and Jörgensen, S.B. (2016) How do entrepreneurs' characteristics influence the benefits from the various
elements of a business incubator? Journal of Small Business and Enterprise Development, 23, 1, 224–239.
Park, J.S. (2005) Opportunity recognition and product innovation in entrepreneurial hi-tech start-ups: a new perspective
and supporting case study. Technovation, 25, 7, 739–752.
Peña, I. (2004) Business incubation centers and new firm growth in the Basque country. Small Business Economics, 22,
Peters, L., Rice, M. and Sundararajan, M. (2004) The role of incubators in the entrepreneurial process. Journal of
Technology Transfer, 29, 1, 83–91.
Phillips, R.G. (2002) Technology business incubators: how effective as technology transfer mechanisms? Technology in
Society, 24, 3, 299–316.
Porter, M.E. (1980) Competitive Strategy. New York: The Free Press.
Powell, W.W., Koput, K.W. and Smith-Doerr, L. (1996) Interorganizational collaboration and the locus of innovation:
networks of learning in biotechnology. Administrative Science Quarterly, 41, 1, 116–145.
Quatraro, F. (2009) Diffusion of regional innovation capabilities: evidence from Italian patent data. Regional Studies, 43,
Rice, M.P. (2002) Co-production of business assistance in business incubators: an exploratory study. Journal of Business
Venturing, 17, 2, 163–187.
Rubin, T.H., Aas, T.H. and Stead, A. (2015) Knowledge flow in technological business incubators: evidence from
Australia and Israel. Technovation, 41-42, 11–24.
Sá, C. and Lee, H. (2012) Science, business, and innovation: understanding networks in technology-based incubators.
R&D Management, 42, 3, 243–253.
Schneider, C. and Veugelers, R. (2010) On young highly innovative companies: why they matter and how (not) to policy
support them. Industrial and Corporate Change, 19, 4, 969–1007.
Schwartz, M. (2013) A control group study of incubators’ impact to promote firm survival. Journal of Technology
Transfer, 38, 3, 302–331.
Schwartz, M. and Hornych, C. (2010) Cooperation patterns of incubator firms and the impact of incubator specialization:
empirical evidence from Germany. Technovation, 30, 9, 485–495.
Scillitoe, J.L. and Chakrabarti, A.K. (2010) The role of incubator interactions in assisting new ventures. Technovation,
30, 3, 155–167.
Sedita, S.R. and Apa, R. (2016) The role of open innovation-oriented strategies in the innovation performance of
mechanical engineering start-up firms in Northern Italy. In: Parrilli, M.D., Fitjar, R.D. and Rodríguez-Pose, A. (eds.),
Innovation Drivers and Regional Innovation Strategies. London and New York: Routledge, pp. 142–160.
Sedita, S.R., Belussi, F. and Grandinetti, R. (2016) How does a networked business incubator fuel cluster emergence? A
theoretical discussion and an empirical illustration. In: Boari, C., Elfring, T. and Molina-Morales, F.X. (eds.),
Entrepreneurship and Cluster Dynamics. London and New York: Routledge, pp. 104–128.
Sherman, H. (1999) Assessing the intervention effectiveness of business incubation programs on new business start-ups.
Journal of Developmental Entrepreneurship, 4, 2, 117–133.
Soetanto, D.P. and Jack, S.L. (2013) Business incubators and the networks of technology-based firms. Journal of
Technology Transfer, 38, 4, 432–453.
Soetanto, D.P. and Jack, S.L. (2016) The impact of university-based incubation support on the innovation strategy of
academic spin-offs. Technovation, 50-51, 25–40.
Sofka, W. and Grimpe, C. (2010) Specialized search and innovation performance–evidence across Europe. R&D
Management, 40, 3, 310–323.
Stinchcombe, A.L. (1965) Social structure and organizations. In: March, J.G. (ed.), Handbook of Organizations.
Chicago: Rand McNally, pp. 142–193.
Sullivan, D.M. and Marvel, M.R. (2011) Knowledge acquisition, network reliance, and early-stage technology venture
outcomes. Journal of Management Studies, 48, 6, 1169–1193.
Sung, T. K. (2007) Incubators and business ventures in Korea: Implications for manpower policy. International Journal
of Technology Management, 38, 3, 248–267.
Tamásy, C. (2007) Rethinking technology-oriented business incubators: developing a robust policy instrument for
entrepreneurship, innovation, and regional development. Growth and Change, 38, 3, 460–473.
Teece, D.J. (1986) Profiting from technological innovation. Research Policy, 15, 6, 285–306.
Totterman, H. and Sten, J. (2005) Start-ups, business incubation and social capital. International Small Business Journal,
23, 5, 487–511.
Trajtenberg, M. (1990) A penny for your quotes: patent citations and the value of innovations. The Rand Journal of
Economics, 21, 172–187.
Vaona, A. and Pianta, M. (2008) Firm size and innovation in European manufacturing, Small Business Economics, 30, 3,
Von Zedtwitz, M. and Grimaldi, R. (2006) Are service profiles incubator-specific? Results from an empirical
investigation in Italy. Journal of Technology Transfer, 31, 4, 459–468.
Wakelin, K. (2001) Productivity growth and R&D expenditure in UK manufacturing firms. Research Policy, 30, 7,
West, J. and Lakhani, K. R. (2008). Getting clear about communities in open innovation. Industry and Innovation, 15, 2,
Yayavaram S, Ahuja G. (2008) Decomposability in knowledge structures and its impact on the usefulness of inventions
and knowledge-base malleability. Administrative Science Quarterly, 53, 2, 333–362.
Appendix A. Additional descriptive statistics
This appendix contains additional descriptive statistics. In Tables A1 and A2, we compare the means, the correlation
structure and the covariance matrix of data for which we collected non-missing responses and data with missing values for
the dependent variable. We fail to reject the null hypothesis of identical populations in all multivariate tests. This means
that 243 observations are representative of the 409 sampled firms. For this reason, we do not expect a non-response bias
within the sample.
INSERT TABLE A1 ABOUT HERE
Table A2 reports the pairwise correlation coefficients between our variables. Coefficients do not show unexpected signs
and significance, supporting the reliability of the dataset.
INSERT TABLE A2 ABOUT HERE
Appendix B. Principal component analysis
Table B1 provides the scoring coefficients used to estimate the firm components, while Table B2 shows the pattern matrix
explaining the relevance of each variable in the factor. Factor BUSINESS is mostly defined by MKTG and MAN, whereas
and factor TECHNICAL is defined by TECH and ICT. The last column of Table B2 gives the fraction of variance that is
“unique” to the variable (i.e., not shared with other variables). For example, 24% of the variance in TECH is not shared
with other variables in the factor model. Finally, Table B3 reports the proportion of total variance explained by our two
principal components. We followed the Kaiser criterion retaining only those factors with eigenvalues equal or higher than
1. We also used an orthogonal rotation to obtain factors that are not correlated to each other. This procedure serves to
create new indexes without inter-correlated components. As it can be seen, the BUSINESS component explains almost
40% of total variance, and another 36% is explained by the TECHNICAL component. According to the Likelihood Ratio
(LR) test, a factor model must be preferred to a perfect-fit model.
INSERT TABLES B1-B3 ABOUT HERE
Appendix C. Survivorship bias
Because our sample consists of firms having survived up to 2012, as is common in survey-based analyses, our
results may suffer from a survivorship bias. Although we are not able to directly address this issue, we can indirectly
assess the impact of the selection process on our estimates.
Given that we study the role of business incubators on firm’s innovative performance, we should consider the
firm survival after the graduation period. Indeed, during the first seven years of life, both innovative and non-innovative
firms face an important selection process (Cesif and Marsili, 2005). To address this issue, we first split our sample into
two groups belonging to different phases of the selection process, and then we investigate whether the selection process
affects both the level of covariates and the estimated coefficients of the two groups.
In order to create two almost equally numerous groups and considering the upper bound of the incubation
period, we divide our sample into young firms (aged 5 years or less) and mature firms (aged 6 or 7 years), and then we
decompose the difference in innovation performance between the two groups, Δ, following the methodology provided in
Bauer and Sinning (2008). Formally, the decomposition equation can be written as:
Δ=𝑆(𝛽∗,𝑋!")−𝑆(𝛽∗,𝑋!")+𝑆𝛽!,𝑋!" −𝑆(𝛽∗,𝑋!")+𝑆𝛽∗,𝑋!" −𝑆(𝛽!,𝑋!"), (C1)
where 𝑆𝛽!,𝑋!" is the conditional expectation of 𝑌
!" evaluated at the parameter vector 𝛽!, 𝛽∗=ωβ!+(1−ω)β!, and
subscripts M and Y denote mature and young firms respectively. In equation (C1), the right-hand side displays the
difference in innovation performance between the two groups due to differences in observable characteristics (the term in
the first square bracket) and the differences in performance due to differences in the estimated coefficients (second and
third square bracket).1 On the one hand, by assuming the same vector of coefficients for both groups, we are able to
evaluate if young and mature firms differ enough in the values of covariates to exhibit different innovation performance.
On the other hand, by assuming the same levels of covariates, we can understand whether unobserved group factors
influence the marginal impacts of our explanatory variables. Since the selection process may change the endowment of
important explanatory variables as well as the marginal impact of these explanatory variables, a decomposition analysis
allows taking into account both effects.
Table C1 shows that young and mature firms do not significantly differ in terms of both endowments of explanatory
variables and estimated coefficients.
INSERT TABLE C1 ABOUT HERE
Table 1. Industries and activities entering the analysis
ATECO 2002 code
ATECO 2002 description
Computer and related activities
Research and development
74.20.1 and 74.20.2
Architectural and engineering activities and related technical consultancy
Manufacture of machinery and equipment n.e.c
Manufacture of electrical machinery and apparatus n.e.c.
Manufacture of motor vehicles, trailers and semi-trailers
35 (excluding 35.1)
Manufacture of other transport equipment
Table 2. Description of variables
Fraction of firm’s turnover from products new to the market
Dummy variable (1/0), assumes value 1 if the firm passed through an
incubator (either public or private) and 0 otherwise.
Number of types of collaborations for innovation (factor variable
ranging from 0 to 5)
Principal component for marketing and managerial capabilities.
Principal component for technological and ICT capabilities.
Number of types of IP protections
Percentage of turnover in R&D in the firm first year.
Dummy variable (1/0), assumes 1 if the firm is located in the North-
Western part of Italy, 0 if the firm is located in the North-Eastern
part of Italy.
Dummy variable (1/0), assumes value 1 if the firm is a KIBS and O
if the firm is a MEF
Years from firm foundation, calculated in 2012, spans from 3 to 7.
Number of active founders and employees in 2012.
Dummy variable (1/0), assumes value 1 if the company is a spin-off
and 0 otherwise.
Dummy variable (1/0), assumes value 1 if the company report
foreign sales and 0 otherwise.
Variables used to construct
BUSINESS and TECHNICAL
Firms evaluation of technological capabilities after three years from
Firms evaluation of marketing capabilities after three years from the
Firms evaluation of management capabilities after three years from
Firms evaluation of ICT capabilities after three years from the
Table 3. Descriptive statistics (243 obs)
Variables used to construct BUSINESS and TECHNICAL
Table 4. Determinants of firms innovation performance (Ordered
Robust standard errors in parentheses. Significant at: *10%,
**5%, ***1%. aInference is based on one-tailed significance test.
Table 5. Determinants of firms innovation performance with interaction
effects (Ordered Logit)
Robust standard errors in parentheses. Significant at: *10%, **5%,
***1%. aInference based on one-tailed significance test.
Table A1. Multivariate test for non-response bias (within sample)
Wilks' lambda F(14, 287)
Pillai's trace F(14, 287)
Lawley-Hotelling trace F(14, 287)
Roy's largest root F(14, 287)
Table A2. Pairwise correlation matrix (*denotes significance at 5% confidence level)
Table B1. Scoring coefficients
Table B2. Pattern matrix
Table B3. Total variance captured by components
LR test: 𝜒!6=305.45 , p-value=0.000.
Table C1. Bauer and Sinning decomposition analysis
1 Notice that, when ω=1, the coefficient effect is estimated by using young firms’ covariates domain, while when ω=0, the coefficient effect is
estimated by using mature firms’ covariates domain.