Channels of academy-industry interaction and benefits from firms and researchers’

perspective: evidence from Mexico

Abstract

1
 Introduction........................................................................................................................................... 2


2
 PRO‐industry
linkages
in
Mexico ........................................................................................................... 4


3
 Research
designs
and
descriptive
statistics........................................................................................... 6


3.1
 Data
collection
and
sample
characteristics .................................................................................... 6


3.2
 The
model
and
estimation
procedures........................................................................................... 8


3.3
 Descriptive
statistics:
benefit
and
channels.................................................................................. 11


4
 Main
findings....................................................................................................................................... 14


4.1
 Estimation
of
Heckman
models
I:
researchers’
data .................................................................... 14


4.2
 Estimation
of
Heckman
models
II:
firms’
data .............................................................................. 17


5
 Conclusions.......................................................................................................................................... 20


 2


1 Introduction1
It is broadly recognised that universities and public research centres are producers and
transmitters of knowledge, and as such can make important contributions to increase firms’
economic performance, and solve national problems. In this sense, the role of universities and
public research centres, hereafter public research organizations (PRO), is evolving from the
formation of human resources and knowledge generation to the more oriented focus of solving
problems and contributing to development. In the case of developing countries, they can promote
economic and social development and contribute towards the satisfaction of social needs
(Arocena and Sutz, 2005). PRO-industry interactions are one of the key elements of the National
System of Innovation (NSI). However, there is a general perception that developing countries
lack strong interactions, which is in the base of the weaknesses of their NSI (Cimoli 2000;
Dutrénit et al, 2010; Lorentzen 2009; Cassiolato, Lastres and Maciel, 2003; Lall and Pietrobelli,
2002; Muchie, Gammeltoft and Lundvall, 2003). Stronger PRO-industry interactions can play a
role to consolidate NSI, as they can promote virtuous circles in the production and diffusion of
knowledge.

Empirical evidence suggest that the process of knowledge transfer between PRO and industry
occurs through multiple channels such as human resources formation, open science, personnel
mobility, informal contacts, consulting relationships, joint and contract research projects,
patenting and spin-offs (Cohen, Nelson and Walsh, 2002; D’Este and Patel, 2007; Laursen and
Salter, 2003). From the industry perspective, some authors argue that information diffusion,
property rights, human resources, collaborative R&D projects, and networking are the most
important channels of knowledge flows (Swann, 2002; Cohen, Nelson and Walsh, 2002; Narin,
et al, 1997). Several studies go further into the analysis and show that knowledge flows are
sector and technology specific; as different sectors have different knowledge bases and
innovation patterns (Pavitt, 1984), they also have different ways to interact with the academy and
other sources of knowledge.2 From the academic perspective, Meyer-Krahmer and Schmoch
(1998) found that collaborative R&D is the most important knowledge flows in some fields.
Bekkers and Bodas Freitas (2008) found that the relative importance of the channels is similar
amongst firms and academic researchers, however academic researchers assign more importance
to the different channels than firms.

Referring to the benefits obtained through PRO-industry interactions, some authors argue that
firms obtain a different perspective for the solution of problems and in some cases perform
product or process innovations that without interaction would not have been possible, they also
benefit from highly skilled research teams, new human resources, and access to different





























































1
We would like to thank Diego Ubfal, University of California, Los Angeles, for providing valuable econometric
assistant and comments.
2
For biotechnology and pharmaceuticals, Cohen, Nelson and Walsh (2002) found that knowledge transfer by
publications is more important. For chemical, different knowledge flows are found important, such as patents
(Levin, 1988), collaborative research and human resources mobility (Schartinger, et al, 2002), and scientific output,
informal contacts and students (Bekkers and Bodas Freitas, 2008). For electronics, the most important is human
resources, especially students (Balconi and Laboranti, 2006; Schartinger, et al, 2002).

 3


approaches for problem-solving (Rosenberg and Nelson, 1994). For researchers, benefits include
obtaining additional funding for the laboratories and exchanging knowledge (Meyer-Krahmer
and Schmoch, 1998), to secure funds for research assistants and laboratory equipment, to gain
insights for their own academic research, to test applications of a theory and to supplement funds
for their own academic research (Lee, 2000), acquire a new perspective to approach industry
problems and the possibility to shape the knowledge that is being produced at the academy.3
Other works have identified some disadvantages of academy-industry interaction. They mention
that a greater involvement with industry can corrupt academic research and teaching, keeping
away the attention from fundamental research. It can destroy the openness of communication
among academic researchers and put restrictions on publishing, which is an essential component
of academic research (Cohen, Nelson and Walsh, 2002; Mansfield, 1990; Monjon and
Waelbroeck, 2003; Welsh et al, 2008; Rosenberg and Nelson, 1994).

Arza (2010) in this Special Issue discusses channels of PRO-industry interactions and their
associated benefits and costs in the case of developing countries, and claims that different
channels of interaction provide different types of benefits for both agents and also a different
balance in the relation between benefits and costs. The paper also proposes a methodology to
analyse the link between channels and benefits.

Policymakers are keen to promote PRO-industry interactions. However, they have hardly
recognised that both agents respond to different incentives, academic researchers function within
an academic logic, while firms’ depend on business logic. In fact, academy and firms interact for
different reasons. In this sense, differences between both perspectives are important to
understand the evolution of PRO-industry interactions and promote specific policies to
strengthen such interactions.

Despite the growing interest regarding PRO-industry linkages, our knowledge about the drivers,
channels of interaction and benefits from interaction is still limited, particularly in the context of
developing countries. This paper analyzes PRO-industry interactions from both perspectives in
the case of Mexico and discusses differences between the effectiveness of different channels in
terms of the benefits perceived by both agents. A central question to explore is which channels of
interactions are the most effective for each agent to increase their benefits. In addition, this paper
discusses the drivers of interaction for both agents.4 The
 cases
 of
 Argentina
 (Arza
 and
 Vazquez,

2010),
 Brazil
 (Fernandes
 et
 al,
 2010),
 and
 Costa
 Rica
 (Orozco and Ruiz, 2010)
 are
 presented
 in
 this

Special
Issue.
 This study is based on original data collected by two surveys carried out in Mexico
during 2008, to R&D and product development managers of firms and to academic researchers.

This paper is divided into five sections. Section 2 describes the context in which we can analyse
PRO-industry interactions in Mexico. Section 3 describes the methodology and data gathering,






























































3
Welsh et al (2008) found that, for the agricultural-biotechnology sector, an important benefit is the increase of the
contacts between researchers and farmers.
4
This study is based on an international research project titled “Interactions between universities and firms:
searching for paths to support the changing role of universities in the South“, developed under the umbrella of the
Catching up project; it was sponsored by the IDRC (Canada). It compares PRO-industry interactions of 12 countries
from Latin America, Asia and Africa. The questionnaires were discussed between all the national teams.

 4


and presents the Heckman model used to analyze the data. Section 4 contains the main findings
and section 5 concludes.

2 PRO-industry linkages in Mexico

The Mexican NSI is not a consolidated system, but rather a system in building process where not
all key actors do in fact exist, and interactions among the actors who do exist are weak, irregular,
and in some cases are absent. The generation, dissemination, and absorption rate of technological
knowledge within and among institutions is low and mainly restricted to academic and public
research institutions. The Mexican NSI shows not only a poor performance in terms of scientific
and technological productivity; also PRO-industry linkages are very fragile.

Higher education in Mexico is rooted to 1910 with the creation of the first university, the
National University of Mexico (UNAM). The education system was consolidated during the
period of import substitution model during 1940s. Other major public and private universities,
such as the National Polytechnic Institute (IPN), the Technological Institute of Higher Studies of
Monterrey (ITESM), the Metropolitan Autonomous University (UAM) and various state
universities were established between 1930 and 1980. Although, the National School of
Engineering was part of the UNAM since 1910, the foundation in 1936 of the IPN, strongly
oriented to engineering and technological research, marked a fundamental turning point in the
policies of the Mexican state which since then, were oriented not only towards the higher
education but to science and technology.

Between 1935-1945 and 1970-1982, almost all public research centres (PRC) and universities
responsible of the science and technology activities in México were created. Both waves of
organization creation coincided with the formulation of economic policies centred on State
intervention and the promotion of industrial development. PRC such as the first national health
institutes, the PRC for agricultural, forest and stockbreeding industries and the medical and
research science academies were founded, some of them linked to state firms like the Mexican
Institute of Petrol (IMP) and the Electrical Research Institute (IIE). Associated with the creation
of the Science and Technology National Council (CONACYT) in 1970, the main PRC emerged.
Nowadays Mexico’s higher education systems consist of universities, technological institutes,
state educational institutions, and normal schools. In 2005, there were 2,807 universities and
other higher education institutions, of which 40% were public and 60%, private. In 2006, public
universities attracted nearly 68% of total undergraduates and 58% of postgraduate students
(Dutrénit et al., 2010).

Regarding scientific research and knowledge production, four public universities have been of
remarkable importance: UNAM, IPN, UAM and the Centre for Research and Advanced Studies
(CINVESTAV), which account for nearly 50% of scientific production in Mexico. Another
important sub-set of public research centres is made up of 27 research institutions grouped in
three main scientific and technological areas: ten are oriented towards mathematics and natural
sciences; eight towards social sciences and humanities; eight dedicated towards innovation and
technological development, and one providing financial support for postgraduate studies. A third
sub-set is composed of research institutes linked to government ministries: such as IMP, IIE, the

 5


Institute for Forestry and Farming Research (INIFAP) and the National Institute for Public
Health (INSP) among others.

The evolution of universities and PRC was moulded by a linear approach to R&D policies. The
predominant idea was that scientific research will bring paired technological development, which
at the same time, creates the conditions for the appropriate technologies to become sources of
new products and processes; funding went to science and limited attention was paid to PRO-
Industry linkages. Private R&D expenditure has been weak in Mexico over time. PRO
concentrate the greatest efforts in science; these institutions have been supported mainly by
budgets provided by the Mexican government. As they concentrate the majority of capabilities,
which are relevant for supporting innovation processes requiring a scientific base, they are called
upon to play a relevant role in the innovative process.

The productive sector has largely acted as an isolated actor within the system. With the exception
of a strong interaction with the government that mainly results from the macroeconomic policy
and some industrial incentives derived from government regulations, there is a clear absence of
regular linkages between the Mexican industry and other economic and social actors such as
PRO. Moreover, given that knowledge is generated fundamentally at PRC and at public
universities, the weakness of the linkages with these actors is one of the main factors thwarting
the development of national innovation capabilities. The majority of these interactions have
taken place in what may be denominated the public triad: CONACYT-PRC-public universities.
The density of interactions occurs among CONACYT-PRC, CONACYT-universities,
universities-PRC. This configuration has been built over the past three decades as a result of a
persisting science, technology and innovation policies based on the conception that innovation is
a linear process. However, there have been some successful cases of PRO-Industry linkages,
particularly in firms from the pharmaceutical, autopart, chemical, biotechnology, software,
cement and machining industries.

Recognizing that knowledge generated in PRO plays an important role in driving firm level
innovations, the Mexican government implemented explicit policies since the early 1990s to
stimulate PRO-industry linkages. They were strengthened at the end of the 1990s, with the
approval of the science and technology Law in 1999 and 2002, and the presentation of the
Special Program for Science and Technology 2001-2006 (PECyT for its name in Spanish). The
main programs to foster PRO-industry interaction are the following: (i) the R&D fiscal
incentives, which in the call referring to 2007 incorporated that projects should have included
activities of linkages with PRO; moreover, 20% of the expenditure was required to be managed
by these institutions; and (ii) Sectoral Science and Technology fund for the economic
development, which operate as a competitive fund where 86.9% of projects were realized in
interaction with PRO.5

As the society and the economic system rapidly advance toward a more intensive production and
exploitation of all types of knowledge, among the different interactions within the innovation





























































5

 
 Other programs included: (i) IDEA (Incorporation of Mexican Scientists and Technologists to the Social and
Productive Sectors of the Country), operating under the scheme of scholarship payment to Mexican professionals
with master or doctorate degree which joins a firm to develop an R&D and innovation project; (ii) Sabbatical stays
inside firms; (iii) Consortia (public-private partnership) and Strategic Alliances and Innovation- AERI (Networks).



 6


system, PRO-Industry linkages have drawn attention as one of the central factors underlying the
innovative process dynamic. However, only few studies have analyzed PRO-industry
interactions in Mexico, most of them based on case studies for specific sectors (Casas, 2001) or
centred on the academic capacities of PRO (Casas and Luna, 1997). Much less has been written
regarding the condition and characteristics of PRO-industry interaction from the firm’s
perspective. As far as we know, there is not a study on the importance of the drivers of PRO-
industry linkages from both perspectives, and the benefits that these two agents of the NSI could
derive from those linkages in Mexico. This study aims to contribute in the understanding of the
relationships between these factors.

3 Research designs and descriptive statistics

3.1 Data collection and sample characteristics

This study is based on original data collected by two surveys on PRO-industry interactions
carried out in Mexico during 2008. Firms´ survey was answered by R&D and product
development managers. Academic’s survey was answered by researchers working at PRO.

The original firms’ database was integrated by 1200 firms; this included firms benefited from
public funds to foster R&D and innovation activities, such as the R&D fiscal incentives program
and competitive funds to foster innovation, and firms that do not get those benefits. The response
rate was 32.3%. The academic researchers’ database was built from researchers listed at the
National Researchers System (NRS) from CONACYT;6 researchers from six fields of knowledge
were included (Physics & Mathematics7; Biology & Chemistry; Medicine & Health Sciences;
Social Sciences; Biotechnology & Agronomy; and Engineering). Some researchers not
recognized by this NRS were also included in the database. 5,880 researchers integrated the final
database. The percentage of response was 8%.

For this specific study the sample was 325 innovative manufacturing firms and 385 researchers
ascribed to PRO.8 From the sample we identified 245 firms that interact with PRO and 235
researchers that interact with firms.

a) Researchers’ characteristics

In our sample 61.0% of the researchers interact with industry. 81% of the total sample is part of
the NRS, 57% of them have links with industry, while 77% of researchers not listed in the NRS





























































6
The National Researchers System was created in 1984 and its main objectives include the promotion of formation,
development and consolidation of a critical mass of high-level researchers, mainly inside the public system. It grants
pecuniary (monthly compensation) and non-pecuniary (status and recognition) incentives to researchers based on
their productivity and quality of their research.
7
This field also includes Earth Sciences.
8

Our sample does not include researchers from social sciences as they have limited links with industry.

 7


have links with firms. Mostly PhD researchers integrate our sample, as 87% of them have a PhD
degree, 7% have a master degree and 6% are graduated. In terms of gender, men mostly integrate
the sample (78%), they also tend to interact more than female (66% and 44% respectively). The
average age in the sample is 47 years old.

In terms of the institutional affiliation, 58% of researchers are ascribed to universities. However,
researchers from PRC tend to connect more than those affiliated to universities (75% and 51%
respectively). In terms of the partner size, 36% of researchers interact with micro/small firms,
26% with large firms and 38% with both micro/small and large firms. 71% of researchers belong
to a research group, and 61% of the research groups have links with firms. The oldest research
group in our sample is 56 years old. In average, the size of research groups is around 18
members (including PhD, master, graduates, technicians and students of different levels, only
some research groups have Post-Docs).

Data about researchers’ outputs shows that researchers with linkages have more applied research
related products (patents and software) than researchers without linkages. However, regarding
papers indexed in ISI, there is not a significant difference in productivity between researchers
that interact and those who do not interact (see Table 1).9

Table
1 Researchers’ characteristics


No linked Linked
National Solicited Patents 12.67% 25.11%
Foreign Solicited Patents 4.67% 6.81%
Type of output
National registered Software 4.67% 10.64%
Reported papers indexed in ISI 91.33% 85.53%
Average publication of Papers
9.20 9.44
indexed in ISI
Productivity
Annual publication of Papers
0.45 0.45
indexed in ISI(a)

b) Firms’ characteristics

75% of firms in our sample have links with PRO (67% of firms interact with universities and
47% with PRC). The sample includes all the manufacturing sectors. We observe that firms from
all sectors tend to connect with PRO, except for the case of Publishing, printing and reproduction
of recorded media; however, the composition between linked and not linked firms differ between
sectors.


In terms of innovation, the sample includes firms that develop innovation of different novelty
degree. 25% introduce improvements in products and/or process, 31% introduce incremental
innovations, and 12% introduce world level innovations. There are some differences between
firms that are linked and those that are not (Table
 2). World level innovators tend to collaborate
more with PRO. In terms of R&D activity, linked firms have larger R&D departments and





























































9
Rivera et al (2009) found the same results in the case of agricultural related research for the case of Mexico.

 8


employ 85% more highly skilled human resources to perform R&D activities than not linked
firms. Firms with links tend to use other information sources more extensively than those
without links.10

Table
2
Firms linkages and innovativeness


Not linked Linked

Novelty of innovations
Number % Number %
Improvements 22 27.5% 22 24.1%
Incremental 23 28.8% 23 31.4%
Radical 4 5.0% 4 13.9%
Total 80 100.0% 80 100.0%

Firms with foreign investment represent 33% of the total sample; they have about the same
tendency to interact than national firms (70%). In terms of firm size, most firms in the sample are
medium-size (42%) and large (42%), only 16% are micro and small. In general we observed that
micro/small and large firms tend to interact more (80%) than medium-size firms (68%). Fiscal
incentives for R&D has been a successful program during the 1990s, 42% of the sampled firms
gets these type incentives, and 84% of them have links with industry.

According to the National Innovation Survey of 2006, 25% of firms are innovators, half of the
innovators perform R&D activities, a tenth of the innovators collaborate with PRO to develop
product and process innovations, and 65% use PRO as information source. Thus, even though
our survey is not statistically significant, its distribution is not far from that observed for
innovative firms.

3.2 The model and estimation procedures

To increase comparability across countries, we followed a model proposed by Arza (2010). The
model suggests that different channels of interactions (Chi) have the potential to trigger different
kinds of benefits, both for researchers (BRi) and for firms (BFi). Moreover, there are some other
researchers’ and firms’ features (Ri and Fi respectively) that may affect their benefits. We follow
two sets of equations, one for researchers and another for firms.

a) Researchers´ perspective:
(1.1) d_V = RViβ + ui


(1.2)
IBi = Chiα + Riδ + εi

(1.3)
d_V = RViβ + ui


(1.4) EBi = Chiα + Riδ + εi

b) Firms´ perspective:





























































10
As most of the studies have found, for both types of firms, customers are the most important type of knowledge
source.

 9


(2.1)
d_V = FViβ + ui


(2.2) PBi = Chiα + Fiδ + εi

(2.3)
d_V = FViβ + ui


(2.4) IBi = Chiα + Fiδ + εi

For the selection part of the model (equations 1.1, 1.3, 2.1 and 2.3), the dependent variable (d_V)
is a dummy variable that equals 1 when firms or researchers interact. Independent variables are
those features of researchers (RVi) and firms (FVi) that affect their probability to interact. The
benefits from interaction, modeled by equations (1.2, 1.4, 2.2 and 2.4), are measured for
researchers and firms who actually interact.11

For the researchers’ and firms’ models we first selected the variables of the selection model that
may affect the probability of linking. Secondly, we identified the best possible model for the
selection equation by estimating different specifications of Probit models on the probability of
linking. To select the variables that better fit the model we performed a log-likelihood ratio test
(LR) on the Probit models. Thirdly, we selected the variables that better describe the benefits
from PRO-industry interaction and we tested them on the overall Heckman model.

a) Researchers

For the selection equation (1.1 and 1.3) we built on the literature on drivers of linkages from the
researchers perspective and identified the variables that better explain the probability of linking.

For equations 1.2 and 1.4 of the Heckman model, benefits are the dependent variable. In our
case, we performed a factor analysis and grouped the benefits in two factors: Intellectual and
Economic.12 Table A.1 in the Annex presents the rotated matrix for benefits and Table 8 below
describes both groups.

The main explanatory variables for equations 1.2 and 1.4 are channels of PRO-industry
interactions, which we categorized ad-hoc in 4 different groups: Bi-directional, Commercial,
Services and Traditional (see Arza, 2010 in this issue).13 Other independent variables for these
equations are associated to knowledge skills, academic collaboration, networking with firms and






























































11
See Arza (2010) in this special issue for a broader discussion of the model construction.
12
The common explained variance by these two factors is 69.89%
13
The industrial parks variable of the Bidirectional channel and the internships of the Services channels had many
missing values in the Mexican database; we decided to eliminate these variables to gain observations in the whole
model. To build the indicator of Channels of PRO-industry interaction we rely on a question where researchers and
firms evaluated the importance of each mode of interaction. This design can lead to multicolinearity among these
variables and standard deviation (see Table A.4 and A.5 in the annex). However we tested each Channel
individually in the overall model, we did not find drastic changes on the coefficients or the significance of other
variables nor the overall model, but the significance of each individual channel increased when tested them
individually.

 10


institutional affiliation. Table 3 describes the main concepts used for measuring the probability of
linking and the determinants of the benefits, and the variables used to proxy them.

Table 3 Variables for analysing PRO-industry linkages from the researchers’ perspective

Eq. Eq.
Characteristic Variable Type of variable
1.1/1.3 1.2/1.4
Degree x x Dummy: PhD=1; Master=1; Graduate=0
Type of Dummy: Basic science=1; Technology
x
research development=1; Applied science=0
Knowledge skills
Dummy: Physic & Mathematics=0; Chemistry &
Research field x x Biology=1; Medicine & Health Sciences=1;
Biotechnology & agronomy=1; Engineering=1
Member of a
x Dummy: Yes=1; No=0
research team
Numerical: RH=ΣxijPi/N
Academic Human
collaboration Postdoc=0.4, PhD=0.4, PhD students=0.3, Master
resources in the x
students and researchers=0.2, students, College
team
researchers and technicians=0.1
Team age x Numerical
Importance of
linking with x Dummy: Yes=1; No=0
Networking with
firms
firms
Initiative of Dummy: Firms’ initiative=1; Both=1; Researchers’
x
collaboration initiative=0
Institutional Type of
x x Dummy: 1=University, 0=PRC
affiliation organization

b) Firms

For the selection equation (2.1 and 2.3) we rely on the literature regarding drivers of linkages
from the firms’ perspective, and identified the variables that better explain the probability of
linking.

For equation 2.2 and 2.4 of the Heckman model, benefits are the dependent variable. Benefits are
defined ad-hoc as Innovative and Productive (see Arza, 2010 in this Special Issue) and Table 5
below. The main explanatory variables for firms’ benefits are related to channels of PRO-
industry interactions. The other independent variables for these equations are associated with:
Innovative capabilities, Firms’ characteristics, and Strategy. Regarding strategy we analyzed two
variables: use of fiscal incentives for R&D and openness strategy. We drew on Laursen and
Salter (2004) to build 4 factors by principal components that express the openness strategy to get
information from external sources:14 Access to open science, consulting and research projects






























































14
Laursen and Salter (2004) argue that management factors, such as firms´ strategy to rely on different types of
information sources, among others, are important drivers to collaborate and get the benefits from academy. They
used the UK innovation survey and as part of their analysis they built a variable that reflects firms’ search strategies.
From a pool of 15 information sources, excluding universities and within the firm, they performed a factor analysis
using principal components; they obtained two factors for openness strategy. Both factors are significant and
positive to explain the use of university knowledge.

 11


with other firms, market, and suppliers.15 Table 4 describes the main concepts used for measuring
the probability of linking and the determinants of the benefits of firms, and the variables used to
proxy them.

Table 4 Variables for analyzing PRO-industry linkages from the firms’ perspective

Eq. Eq.
Characteristic Variable Type of variable
1.1/1.3 1.2/1.4
Human resources in Numerical: Human resources in R&D as %
x
Innovative R&D of the total employment
capabilities Formalization of R&D Dummy: Formal and continuous innovative
x x
and innovative activities activities=1; Otherwise=0
Firm size x x Numerical: ln of firms’ employees
Categorical: 0.25: low, 0.5: medium-low,
Firms’ Technology sector x x
0.75: medium-high, 1:high
characteristics
Dummy: Foreign investment=1;
Ownership x x
Otherwise=0
Factors loads for: Access to open
information, Consulting and research
Openness strategy F1-F4 x x
projects with other firms, Market, Suppliers.
Strategy
(See table A.4 in the Annex)
Fiscal incentives for
x x Dummy: Yes=1; No=0
R&D
Categorical: 0.25: without importance, 0.5:
Creation and transfer of
Role of PRO x low importance, 0.75: medium importance,
knowledge
1: high importance.

3.3 Descriptive statistics: benefit and channels

a) Benefits from interaction

Regarding the benefits associated to interaction from the researchers’ perspective, Table 5
presents the average of importance of each variable individually and for Economic and
Intellectual benefits.

Table 5 Importance of benefits for researchers

Benefits Average (scale 0.25-1)

Economic benefits 0.56
Share equipment / instruments 0.50
Provision of research inputs 0.56
Financial resources 0.61

Intellectuals benefits 0.69

Ideas for further collaboration projects 0.73
Inspiration for further scientific research 0.70
Share of knowledge/information 0.66





























































15
The common explained variance by these two factors is 66.1%. See Table A.4 in the Annex for a better
description of the factor analysis.

 12


Reputation 0.65

Researchers rank higher Intellectual benefits than Economic benefits. This suggests that
researchers are knowledge driven rather than economic driven. The most important individual
benefits for researchers are related to new collaboration projects and new scientific research.
Regarding Economic benefits, researchers rank higher financial resources for research, which is
linked to the development of new projects. We found that Biotechnology and Agronomy report
higher Economic and Intellectual benefits than other research fields.

Regarding the benefits associated to interaction from the firms’ perspective, Table 6 presents the
average of importance of each variable individually and for Innovative and Productive benefits.
For the case of firms, Productive benefits are more important than Innovative benefits, which
suggest that firms from our sample are more interested in solving short-term problems. The most
important individual benefit is associated with contact students for future recruiting. Regarding
innovative benefits, the most important individual benefit is associated with absorbing
technological information. Firms in low technology sectors rank higher Productive benefits,
while firms located in low and medium-high technology sectors rank higher Innovative benefits.


Table 6 Importance of different types of benefits for firms

Benefits Average (scale 0.25-1)

Innovative benefits 0.39
Technology transfer from the university 0.39
To augment the firms ability to find and absorb
0.40
technological information
To get information about trends in R&D in the field 0.37
To contract research to contribute to the firms’
0.39
innovative activities
To contract research that the firm do not perform 0.38
Productive benefits 0.42
To get technological/consulting advice from
0.43
researchers to solve production/related problems
To make earlier contact with excellent university
0.47
students for future recruiting
To use resources available at universities and PRC 0.40
To perform test for products/processes 0.44
To help in quality control 0.38

b) Channels for PRO-industry interaction

Regarding the channels of interaction, Table 7 shows the average of importance and the percent of
higher importance for each mode and channel of interaction for researchers and firms.

Researchers and firms have different perceptions regarding the importance of channels of PRO-
industry interaction. Researchers value more the Bi-directional channel (60%), particularly
knowledge transferred from cooperative research. On the other hand, firms value more the

 13


Traditional channel (58%). This mismatch suggests that from the firms´ perception, PRO above
all contribute with the creation of key human resources and the diffusion of knowledge, while for
the researchers’ perspective, the generation of knowledge is a key role. There is an agreement
regarding the Commercial Channel, for both is the least important.

Table 7 Importance of Channels and forms of PRO-industry interaction

Researchers’
Firms’ perspective
perspective
% of
% of firms
Channels and Forms of interaction researchers
for whom
Average for whom Average
it is
it is
important
important
Traditional 0.54 37.7 0.58 47.7
Publications 0.50 30.1 0.59 45.3
Conferences and expos 0.61 48.6 0.58 48.9
Graduates employed recently 0.53 34.3 0.57 48.9
Services 0.58 47.3 0.54 40.0
Consultancy 0.60 50.1 0.54 40.3
Internships (mobility) 0.48 32.7 0.45 25.2
Informal information exchange 0.65 57.7 0.56 41.9
Training staff 0.59 48.8 0.61 52.6
Bidirectional 0.60 49.0 0.54 39.6
Research contracts 0.64 55.3 0.54 37.8
Cooperative R&D 0.68 61.0 0.58 46.5
Networking 0.58 47.0 0.49 34.5
Commercial 0.48 30.3 0.43 24.8
Spin-off 0.45 25.7 0.34 10.8
Incubators 0.51 35.1 0.44 24.3
Technology licenses 0.47 29.9 0.48 30.8
Patents 0.48 30.6 0.49 33.5




c) Channels and benefits


Regarding the importance of channels of PRO-industry interaction and benefits, Table 8 shows the
different perceptions regarding channels of interaction and benefits for both perspectives. In
general terms, researchers use the four channels more effectively and get more benefits from
interaction. Researchers use more actively Services and Traditional channels, and they usually
report higher intellectual benefits from each channel of interaction. On the other hand, firms use
more actively Bi-directional and Commercial channels. Firms report higher innovative benefits
from the different types of collaboration.

Table 8 Channels and benefit of researchers and firms

Researchers Firms
Forms Importance
Intellectual Economics Innovative Productive
Bi-directional Not important 0.67 0.54 0.49 0.43

 14


(VB) Important 0.74 0.60 0.73 0.66

Commercial Not important 0.68 0.55 0.52 0.46

(VC)
Important 0.73 0.61 0.72 0.69

Not important 0.64 0.52 0.49 0.43

Services
(VS)
Important 0.81 0.66 0.70 0.64

Traditional Not important 0.64 0.52 0.47 0.42

(VT) Important 0.78 0.65 0.67 0.60



4 Main findings



4.1 Estimation of Heckman models I: researchers’ data

Table 9 presents the results of the Heckman model estimated with robust standard errors for
equations (1.1) and (1.2) on intellectual benefits and (1.3) and (1.4) for economic benefits.
Equations (1.1) and (1.3) are estimations for the selection equation, which estimates the
probability that researchers link to firms. These estimations are used as a tool to correct for
selection bias in the estimation of benefits. Results form the selection equations (intellectual and
economic) are fairly similar, which increases the robustness of our model.

Researchers with Master and PhD degrees tend to connect less likely than graduates, as
suggested by the significant, negative and high coefficient of both degrees.16 This result suggests
that the type of links with the industry does not depend on the skill of postgraduate researchers,
because they could be more oriented towards solving simple production problems than to
experimenting in new areas. This result is different from the positive impact of the academic
status found by other authors, such as D’Este and Patel (2007) and Boardman and Ponomariov
(2009), amongst others.

The results of the research fields confirm that there are significant differences between fields as
referred to the likelihood to connect. The significant, positive and high coefficient of both
Biotechnology & Agronomy and Engineering reveal that researchers from these fields tend to
connect much more with industry than researchers from Physics & Maths, as it was expected.
However, in the Mexican case, researchers of Physics & Maths are not less likely to be
connected, in fact Medicine & Health Sciences tend to connect much less. Regarding the type of
research, researchers that carry out basic science and technological development tend to connect
more than those that carry out applied research. The results of the research fields and type of
research confirm that in the Mexican case, fields associated mostly with the quest of fundamental





























































16
To proxy degree we built dummy variables, which equals one if the researcher has a master or PhD degree, and
cero if the researcher has a graduate degree.

 15


understanding, as Physics & Maths are more likely to connect than other fields that are seen as
more problem-oriented, like Chemistry & Biology and Medicine & Health Sciences. The low
public funding of basic science may contribute to explain why these researchers look for other
funding sources, such as firms, to complement the funding of their research activities. However,
we have to notice that we cannot claim that this is the behaviour of researchers in Mexico, as our
sample is not statistically representative.

Table 9 Heckman estimates of economic and intellectual benefits for researchers

Intellectual
Selection Selection Economic Benefit
Benefits
(1.1) (1.3) (1.4)
(1.2)
-0.6401** 0.5444* -0.7716** -0.2689
Master
(0.3162) (0.3192) (0.3109) (0.3592)
-1.2633*** 0.6630*** -0.9215*** 0.0571
PhD
(0.2603) (0.2310) (0.2360) (0.2706)
0.1999 -0.1885 0.2099 0.0776
Chemistry & Biology
(0.1812) (0.2664) (0.1919) (0.2294)
Medicine & Health -0.6124** -0.1942 -0.3529 0.1564
Sciences (0.2921) (0.4322) (0.2612) (0.3800)
Biotechnology & 1.1861*** -0.2436 1.0305*** -0.1869
Agronomy (0.1800) (0.2244) (0.2014) (0.2246)
0.4770*** -0.3317 0.5216*** -0.2156
Engineering
(0.1653) (0.2260) (0.1629) (0.2079)
0.5543*** 0.4924**
Basic science
(0.1379) (0.2108)
0.8822*** 0.6772**
Technology development
(0.1682) (0.3355)
0.4668*** -0.0539
Member of a research team
(0.1376) (0.2041)
-0.0087* 0.0116**
Team age
(0.0053) (0.0058)
Human resources in the 0.0062* -0.0081**
team (0.0029) (0.0030)
-0.5716*** 0.1366 -0.4057*** 0.0838
Type of organization
(0.1240) (0.1181) (0.1166) (0.1345)
Importance of linking with 1.6300*** 1.5660***
firms (0.1131) (0.0981)
Firms initiative of -0.2751* 0.1713
collaboration (0.1593) (0.1967)
Both initiative of -0.2182 0.0623
collaboration (0.1175) (0.1273)
0.8433* 0.1534
Traditional
(0.4501) (0.3618)
0.7082** 0.1725
Bidirectional
(0.3578) (0.3501)
0.4699 0.1718
Services
(0.3798) (0.4211)
Commercial -1.0629*** 0.0039

 16


(0.3180) (0.3274)
-0.0382 -0.6854* 0.0152 0.0173
_cons
(0.3067) (0.3547) (0.4724) (0.4196)
Observations 382 382
Censored 150 150
Wald Chi2(15) 58.61 36.70
Prob>chi2 0.0000 0.0014
athrho -0.8511*** -1.5601
lnsigma -0.0807 0.0473
Wald test of indep. eqns.
37.48 11.31
(rho = 0): chi2(1) =
rho -0.6916 -0.9154
sigma 0.9225 1.0485
lambda -0.6380 -0.9598
*p ‹ 0.1; **p ‹ 0.05; ***p ‹ 0.005

In relation to academic collaboration, a researcher member of a team is more likely to connect

with industry to obtain intellectual benefits, in other words there is a positive relationship
between academic collaboration and the likelihood to connect with industry. This also highlights
the importance of teamwork when dealing with industrial problems. In contrast, Boardman and
Ponomariov (2009) found that academic collaboration is not a driver of PRO-industry
interaction.

The institutional affiliation also matters; researchers working in PRC are more likely to connect
with industry than researchers working in universities, as found by Boardman and Ponomariov
(2009). In the Mexican case there is a set of explanations for this result. Internal incentives for
research and the stimulus for collaborating with firms differ between universities and PRC.
Universities have more institutional funding for research than PRC and valuate less the
interaction with firms. In addition, more researchers from universities than from PRC are
members of the NSR. Although several reforms have increased the importance of products other
than publication –e.g. patents, technological developments- for the allocation of the stimulus, the
principle of “publish or perish” remains as the main incentive for the NSR. Finally, there is a
significant and positive relationship between the perception about the importance of linkages
with industry and the likelihood to interact with industry.

Equations 1.2 y 1.4 show the outcome of the interaction with industry. We found a positive
relationship between holding a Master and a PhD degree and obtaining intellectual benefits, the
significance is very high in the case of PhD. Although having a Master or a PhD is not a driver to
collaborate, having a postgraduate degree allows researchers to get more intellectual benefits
than having a first degree. In contrast, postgraduate degrees are not important for obtaining more
economic benefits.

Referring to academic collaboration, we found that working in a more robust research team (with
higher degrees) allows obtaining more intellectual benefits than working individually, but as the
research team is more robust the researchers obtain less economic benefits. These results suggest
that firms are more interested in collaborating with graduates to solve simple problems than with
highly formed teams. However, we also found that as the team is more experienced in terms of

 17


years, more economic benefits can be obtained. This suggests that experience in collaborative
projects, allow the team to know how to establish and manage new collaborative projects to get
more economic benefits. In contrast, experience of the team does not contribute to obtain more
intellectual benefits. Who takes the initiative to collaborate has important impacts on benefits. As
the University takes the initiative, it is more likely to obtain intellectual benefits than if firms
take the initiative.

We found that there is a significant and positive relationship between Traditional and
Bidirectional channels, and intellectual benefits. In contrast none of the channels contribute to
receiving economic benefits. For intellectual benefits Bidirectional is more significant than
Traditional. Traditional is related to open science and hiring graduates by the industry; these
forms of interaction do not require formal linkages but allow informal knowledge flows. The
Bidirectional channel includes two forms of interaction that are related to formal interactions,
such as cooperative R&D projects and contract research, we also found that skilled researchers
(PhD degree) are correlated to Bidirectional channels. Thus, data shows that Mexican researchers
that collaborate with industry receive intellectual benefits through formal and informal channels.
In contrast, we found a significant, negative and high coefficient of the Commercial channel,
which includes patents used by firms, technology licenses, incubators and spin-offs. This channel
is the least important of the four channels in the Mexican case (see table 9 in section 3.3). If we
look to patents and technology licences, they have a double face, on one side protect the
knowledge generated through interaction, on the other side, they are a way to diffuse it with
some lag. The negative relationship suggests that for the Mexican researchers the restriction to
knowledge sharing is more important than the possibility to use this knowledge for future
research.

4.2 Estimation of Heckman models II: firms’ data

presents the results of the Heckman model estimated with robust standard errors for
Table
 10
equations (2.1) and (2.2) on production benefits and (2.3) and (2.4) for innovation benefits.
Equations (2.1) and (2.3) are estimations for the selection equation, which estimates the
probability that firms link to PRO. These estimations are used as a tool to correct for selection
bias in the estimation of benefits. Results form the selection equations (productive and
innovative) are fairly similar, which increases the robustness of our model.

 18


Table
10
Heckman
estimates
of
production
and
innovation
benefits
for
firms


Productive
Selection Selection Innovative Benefits
Benefits
(2.1) (2.3) (2.4)
(2.2)
0.0022** 0.0025**
Human resources in R&D
(0.0009) (0.0010)
Formalization of R&D and -0.02810 0.1071** -0.1225 0.0785
innovative activities (.3025) (0.0531) (0.3441) (0.0578)
-0.0022 -0.0022 0.0198 -0.0044
Firm size
(0.0651) (0.0081) (0.0603) (0.0088)
0.2237 -0.0520 0.3484 -0.0314
Technology sector
(0.3555) (0.0508) (0.3581) (0.0506)
0.0603 -0.0355 0.0412 0.0113
Ownership
(0.1897) (0.0281) (0.2022) (0.0307)
0.2323*** -0.0391*** 0.2265** -0.0074
Openness strategy F1
(0.0870) (0.0145) (0.0974) (0.0155)
0.1400 0.0186 0.1977** 0.0337**
Openness strategy F2
(0.0956) (0.0135) (0.0934) (0.0164)
0.0054 -0.0150 0.0448 -0.0186
Openness strategy F3
(0.0844) (0.0129) (0.0825) (0.0123)
0.2066** -0.0086 0.2145** -0.0111
Openness strategy F4
(0.0933) (0.0140) (0.0888) (0.0151)
0.5060** -0.0643** 0.3832** -0.0570*
Fiscal incentives for R&D
(0.1977) (0.0301) (0.1887) (0.0336)
Creation and transfer of 1.1623*** 1.1459***
knowledge (0.3024) (0.3368)
Traditional 0.1330* 0.0146
(0.0731) (0.0706)
Bidirectional 0.2303** 0.2049**
(0.0892) (0.0957)
Services 0.1839* 0.1986*
(0.0963) (0.1029)
Commercial -0.0243 0.0585
(0.0949) (0.1142)
-0.3876 0.2949 -0.4141 0.2637
_cons
(0.4970) (0.0736) (0.5871) (0.0892)
Observations 310 310
Censored obs 69 69
Wald Chi2(14) 174.74 109.51
Prob>chi2 0.000 0.000
athrho -1.0954 -0.6492
lnsigma -1.5909 -1.6105
Wald test of indep. eqns.
12.38 2.03
(rho = 0): chi2(1) =
rho -0.799 -0.571
sigma 0.204 0.200
lambda -0.163 -0.114
*p ‹ 0.1; **p ‹ 0.05; ***p ‹ 0.005

 19


To proxy innovative capabilities we tried different indicators such as the ratio of human
resources in R&D as a percentage of total employees, and the formalization of R&D activities.
But, none of the coefficients was significant, which can be explained by the fact that most of the
firms in our sample perform R&D activities, which generates a bias in the model expressed by
the low significance levels of these variables. However, these two variables have an important
impact to obtain productive benefits; and human resources in R&D has an important effect to get
innovative benefits. Regarding innovative capabilities we reported a combination of indicators
that better explain the behaviour of the firms for selection and benefits; however, it could be
worthy to note that when an indicator of product and process innovations was introduced to the
model, no significant coefficients were found neither for selection nor benefits, which can also
be explained by the bias related to the fact that most of the firms in our sample perform
innovative activities.

Contrary to the results from previous studies (Cohen et al, 2002; Motohashi, 2005; Tether and
Tajar 2008), firms characteristics’ expressed by firm size, technological level of the firm operate
and ownership are not important drivers for interaction with PRO. Productive and innovative
benefits are not affected by these firms’ features either.

Regarding the openness strategy, the results suggests that firms that deliberately search for
external knowledge sources are more likely to establish linkages with PRO than those that do not
follow an openness strategy These results are similar to those found by Laursen & Salter (2004).
Some of those strategies have a higher effect on the firms’ probability to interact with PRO. The
highest effect refers to access to open sources, and the interaction with suppliers, as shown by
highly significant and positive coefficients. The one with less significance is the openness
strategy related to consultancy and research projects with other firms. In contrast, the openness
strategy related to market has no effect on the likelihood of PRO-I linkages.

Another indicator of firm’s strategy is the use of fiscal incentives for R&D. As it is suggested by
the significant, positive and high coefficients obtained for the selection equations (2.1 and 2.3),
firms receiving fiscal incentives for R&D are more likely to connect with PRO than those who
are not getting them. Firm’s perception about the role of PRO for the creation and transfer of
knowledge is also an important driver of linkages.

Referring to the productive and innovative benefits, we found a positive and significant
relationship between human resources in R&D and productive and innovative benefits. We also
found a positive relationship between formalization of R&D and innovative activities and the
achievement of productive benefits. These results suggest that firms’ innovative capabilities
allow firms to gain productive and innovative benefits from interaction.

Even though firms accessing information mainly from open science are more likely to connect to
PRO, we found a negative relationship between this openness strategy and productive benefits
that firms can obtain. This result suggests the existence of a substitution effect i.e. increasing
access to open science allows firms to gradually develop capabilities to solve production related
problems that probably were solved with the assistance of PRO.

 20


We found that fiscal incentives for R&D are a driver of linkages with PRO; however, fiscal
incentives are negatively related with the firms´ production and innovation benefits. These
apparently contradictory results could reflect the fact that fiscal incentives for R&D have just
triggered firms linkages with PRO, but they are not very important in determining benefits.
Furthermore, having connections with PRO increases the firms’ chances of being selected as
beneficiaries of that policy instrument, thus firms make connections with PRO just to fulfil that
requirement, but they do not consciously look for productive or innovative benefits derived from
those connections. The results suggest that fiscal incentives for R&D could play an important
role to push the establishment of PRO-industry linkages, helping to overcome the barrier that
firms hold and contributing to change firms’ behaviour in the future.

Regarding channels of interaction, firms obtain both, productive and innovative benefits from the
Services channel and Bidirectional channel. Bidirectional is the most significant channel in
explaining productive an innovative benefits. A firm could expect higher innovative benefits
from cooperative R&D projects, research contracts and networking. We also found an important
correlation between firms’ innovative capabilities and this channel. The positive relationship
found between the Traditional channel and productive benefits suggests that training of human
resources and diffusing information by PRO play an important role to improve firms´ production
activities.

5 Conclusions
This paper has analyzed PRO-industry interactions from the perspectives of innovative firms and
academic researchers in the case of Mexico. A high percentage of researchers and firms in our
sample are linked, thus it gives quite accurate insights from the perspective of firms and
researchers that actually interact. We focus on analysing differences between the effectiveness of
different channels of PRO-industry interaction in terms of the benefits perceived by both agents.
A central question to explore has been which channels of interaction are the most effective for
each agent to increase their benefits. In addition this paper discusses the drivers of interactions.

Referring to drivers of PRO-industry interaction, we found that the main drivers for firms are:
openness strategy (particularly open science and suppliers), fiscal incentives for R&D,
perception about the role of PRO for the creation and transfer of knowledge. In the case of
researchers, the main drivers are: degree (graduates), research field and type of research
(Biotechnology & Agronomy, and Engineering, but in our sample to a lesser extent Physics &
Maths), academic collaboration, institutional affiliation and perception about the importance of
linking.

Referring to channels, as found by other authors (Schartinger et al., 2002; Bekkers and Bodas
Freitas, 2008), in the Mexican case both agents use a variety of channels for knowledge
interaction. This result confirms previous findings for other authors about Mexico (Casas, 2001).
As it was found in many studies, human resources formation, the creation of new physical
facilities, consultancy, contract and joint research, training, meetings and conferences, are more
important than patenting and spin-offs (Cohen, Nelson and Walsh, 2002; D’Este y Patel, 2007).

 21


Our findings suggest that researchers are knowledge driven rather than economic driven, as they
valuate more Intellectual than Economic benefits. In the case of firms, Productive benefits are
more important than Innovative benefits, which suggest that firms are more interested in solving
short-term problems than in experimenting and innovating.

Referring to the channels of PRO-industry interaction, the Bi-directional channel is more

important for researchers than for firms. These finding reject those by Bekkers y Bodas Freitas
(2008), as they did not find differences among the relative importance of modes of PRO-industry
interaction amongst both partners.

Referring to the channels PRO-industry interaction, the Bi-directional channel is more important
for researchers and for firms. These finding confirms those by Bekkers y Bodas Freitas (2008),
as they did not find differences among the relative importance of modes of PRO-industry
interaction amongst both partners.

Referring to the hypothesis proposed by Arza (2010) in this Special Issue, our findings allow us
to claim that:
• Traditional channels will predominantly provide intellectual benefits for PRO, and
solutions to productive problems for firms.
• Bi-directional channels will predominantly provide intellectual benefits for PRO, insights
for innovative activities for firms but also solutions for production problems, the dual
benefits regarding this channel contributes to building virtual circles for PRO-industry
interaction.
• Services channels will predominantly provide solutions to productive problems and
innovative benefits for firms, but not any benefit for researchers.
• Commercial channels will negatively affect intellectual benefits for PRO.
• Intellectual benefits for PRO are driven predominantly by Bi-directional channels.
• Researchers’ benefits will be larger when the channels are used by researchers with
postgraduate studies, particularly PhD.
• Innovative benefits for firms are driven predominantly by Bi-directional channels.
• Productive benefits for firms are driven predominantly by Bidirectional and Services
channels.

Our findings have some policy implications. The significance of the drivers related to
perceptions about the partner both from the firms and the researchers perspective suggest that
working on the agents perceptions may have an impact on the overall PRO-industry linkages.
The importance of the graduates recently hired from the firms’ perspective suggests that they
could be seen as an important interface between researchers and firms. This call for new policies
oriented to work with new graduates to foster interactions and innovation by the firms. Finally, it
is clear that recent public program to foster PRO-industry interaction, like fiscal incentives for
R&D, have had impact on the interactions building. However, the contradictory results -positive
as drivers of interactions but negative to obtain benefits- suggest that this type of programs could
play an important role to push the establishment of PRO-I linkages, thus helping to overcome the
barrier that firms hold and contributing to change firms’ behaviour in the future.

 22


Annex
Table A.1 Researchers’ benefits. Rotated Component Matrix

Intellectual Economic
benefits benefits
Further collaboration projects 0.900 0.184
Ideas for further research 0.802 0.352
Knowledge/information sharing 0.754 0.324
Reputation 0.653 0.408
Share equipment/instruments 0.319 0.696
Provision of research inputs 0.320 0.803
Financial resources 0.216 0.797
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
Rotation converged in 3 iterations.
Explained variance: 69.89%

Table A.2 Correlation matrix. Channels of PRO-industry interaction. Researchers

Bi-directional Commercial Services Traditional

(VB) (VC) (VS) (VT)
Bi-directional
1
(VB)
Commercial
0.6948 1
(VC)
Services
0.6953 0.644 1
(VS)
Traditional
0.6358 0.5688 0.6977 1
(VT)

Table A.3 Correlation matrix. Channels of PRO-industry interaction. Firms

Bi-directional Commercial Services Traditional

(VB) (VC) (VS) (VT)
Bi-directional 1
(VB)
Commercial 0.8082 1
(VC)
Services 0.8389 0.742 1
(VS)
Traditional 0.7193 0.6468 0.7519 1
(VT)

 23


Table A.4 Firms’ openness strategy. Rotated Component Matrix.

Consulting
Access to and research
Linkages Market Suppliers
open science projects with
other firms
Suppliers .183 .142 .076 .911
Customers .061 .024 .876 .137
Competitors .433 .182 .509 -.226
Joint or cooperative projects with .114 .626 .365 .165
other firms
Consultancy with R&D firms .016 .849 -.076 .059
Publications and technical reports .603 .449 .090 -.095
Expos .693 -.088 .204 .119
Internet .773 .090 -.011 .222
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
Rotation converged in 3 iterations.
Explained variance: 66.1%

References

Arocena, R. and J. Sutz (2005), “Latin American Universities: From an original revolution to an uncertain
transition”, Higher Education, 50, 573-592.
Arza, V. and López, A. (2008), “Characteristics of university-industry linkages in the Argentinean
Industrial Sector”, paper presented at the Globelics Conference, Mexico City.
Balconi M. and A. Laboranti, (2006), “University–industry interactions in applied research: the case of
microelectronics”, Research Policy, 35, 1616–1630.
Bekkers, R., and I.M. Bodas Freitas (2008), “Analysing knowledge transfer channels between universities
and industry: To what degree do sectors also matter?”, Research Policy, 37, 1837-1853.
Boardman, C. and B. Ponomariov (2009), “University researchers working with private companies”
Technovation, 29, 142–153.
Casas, R and M. Luna (Coord.) (1997), Gobierno, Academia y Empresas en México: Hacia una nueva
configuración de relaciones. Plaza y Valdés: Mexico.
Casas, R. (Coord) (2001), La formación de redes de conocimiento, Anthropos: México.
Cassiolato, J.E., H.M.M. Lastres and M.L. Maciel (ed) (2003), Systems of Innovation and Development.
Evidence from Brazil, Edward Elgar: Cheltenham.
Cimoli, M. (ed.) (2000), Developing Innovation Systems, Mexico in the Global Context, London: Pinter
Publisher.
Cohen, W., R. Nelson and J. Walsh (2002), “Links and Impacts: The influence of public research on
industrial RandD”, Management Science, 48 (1), 1-23.
D’Este, P. and P. Patel (2007), “University–industry linkages in the UK: What are the factors underlying
the variety of interactions with industry?”, Research Policy 36 1295–1313
Dutrénit, G. and C. De Fuentes (2009), “Academy-Industry interactions at three different stages of the
linking process: Micro evidence from the perspective of both agents”, paper presented at the
Globelics Conference, Dakar.

 24


Dutrénit, G., M. Capdevielle, J.M. Corona Alcantar, M. Puchet Anyul, F. Santiago y A.O. Vera-Cruz
(2010), El sistema nacional de innovación mexicano: estructuras, políticas, desempeño y
desafíos, UAM/Textual S.A.
Intarakumnerd, P. D. Schiller (2009), “University-Industry Linkages in Thailand: Successes, Failures and
Lessons Learned for Other Developing Countries”, Seoul Journal of Economics Vol. 22 (4).
Joseph, K.J. and Vinoj Abraham (2009), “University–Industry Interactions and Innovation in India:
Patterns, Determinants and Effects in Select Industries”, Seoul Journal of Economics Vol. 22 (4).
Lall, S. and C. Pietrobelli (2002), Failing to Compete. Technology development and technology systems
in Africa, Edward Elgar.
Laursen, K. and A. Salter (2004), “Searching high and low: what types of firms use universities as a
source of innovation?”, Research Policy 33, 1201-1215,
Lee, Y.S. (2000), “The sustainability of university–industry research collaboration: an empirical
assessment”, Journal of Technology Transfer, 25, 111–133.
Levin, R.C. (1988), “Appropriability, R&D spending and technological performance”, The American
Economic Review, 78 (2), 424–428.
Lorentzen, J. (2009), “Learning by firms: the black box of South Africa’s Innovation System”, Science
and Public Policy 36, 1, pp. 33-45.
Mansfield, E. (1991), “Academic research and industrial innovation”, Research Policy, 20, 1-12.
Meyer-Krahmer, F. and Schmoch, U. (1998), “Science-based technologies university–industry
interactions in four fields”, Research Policy, (27), 835–852.
Monjon, S. and P. Waelbroeck (2003), “Assessing spillovers from universities to firms: Evidence from
French firm-level data”, International Journal of Industrial Organization, 21, 1255-1270.
Motohashi, K., (2005). University-Industry collaborations in Japan: The role of new technology-based
firms in transforming the National Innovation System. Research Policy, 34, 583-594.
Muchie, M., P. Gammeltoft and B.A. Lundvall (eds) (2003), Putting Africa First. The making of African
innovation systems, Aalborg University Press.
Narin, F., K.S. Hamilton and D. Olivastro, (1997), “The increasing linkage between U.S. technology and
public science, Research Policy 26 pp. 317–330.
Pavitt, K. (1984). Sectoral patterns of technological change: toward a taxonomy and a theory. Research
Policy, 13, 343-373.
Rasiah, R. (2009), “University-Industry Collaboration in the Automotive, Biotechnology and Electronics
Firms in Malaysia”, Seoul Journal of Economics, Vol. 22 (4).
Rosenberg, N. and R. Nelson (1994), “American universities and technical advance in industry”,
Research Policy, (23), 323-348.
Schartinger, D., C. Rammera, M.M. Fischer and J. Fröhlich (2002), “Knowledge interactions between
universities and industry in Austria: sectoral patterns and determinants”, Research Policy 31,
303–328.
Swann, G. (2002), “Innovative Businesses and the Science and Technology Base: An Analysis Using
CIS3 Data Report for the Department of Trade and Industry”, DTI, London, UK.
Tether, B.S. and A. Tajar (2008), Beyond industry-university links: Sourcing knowledge for innovation
from consultants, private research organizations and the public science-base, Research Policy 37,
1079-1095.
Welsh, R., L. Glenna, W. Lacy and D. Biscotti (2008), “Close enough but not too far: Assessing the
effects of university-industry research relationships and the rise of academic capitalism”,
Research Policy, 37, 1854-1864.