Content uploaded by Anders Broström
Author content
All content in this area was uploaded by Anders Broström on Oct 06, 2022
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=gein20
Download by: [Kungliga Tekniska Hogskola] Date: 06 November 2017, At: 09:59
Economics of Innovation and New Technology
ISSN: 1043-8599 (Print) 1476-8364 (Online) Journal homepage: http://www.tandfonline.com/loi/gein20
Mapping research on R&D, innovation and
productivity: a study of an academic endeavour
Anders Broström & Staffan Karlsson
To cite this article: Anders Broström & Staffan Karlsson (2017) Mapping research on R&D,
innovation and productivity: a study of an academic endeavour, Economics of Innovation and New
Technology, 26:1-2, 6-20, DOI: 10.1080/10438599.2016.1202519
To link to this article: http://dx.doi.org/10.1080/10438599.2016.1202519
Published online: 26 Jul 2016.
Submit your article to this journal
Article views: 224
View related articles
View Crossmark data
Citing articles: 2 View citing articles
Mapping research on R&D, innovation and productivity:
a study of an academic endeavour
Anders Broström
a
and Staffan Karlsson
a,b
a
KTH the Royal Institute of Technology, Stockholm, Sweden;
b
Swedish Research Council, Stockholm, Sweden
ABSTRACT
This paper accounts for the development of the academic endeavour to
determine the firm-level relationship between investments in R&D and
productivity. The impact of 28 highly cited publications within this line
of study is investigated using a combination of bibliometric techniques
and citation function analysis. We show how the attention paid to this
line of research broadens and deepens in parallel to the diffusion of
innovation as a research theme during 2000s. Our findings also suggest
that the attraction of scholarly attention is driven by combination of
broadening interest in the central research question under study and
boundary-pushing methodological contributions made in the key
contributions.
ARTICLE HISTORY
Received 26 June 2015
Accepted 12 May 2016
KEYWORDS
Innovation; productivity;
R&D; citation analysis;
bibliometric analysis
JEL classification
B21; B23; B41; C38; D24
1. Introduction
Following a surge of interest in innovation and its consequences, academic researchers came to pay
significant attention to the question of to what extent and under what conditions firms’investments
in research and development (R&D) activities increase their productivity by means of successful inno-
vation. Reviews of this literature are found in Hall, Mairesse, and Mohnen (2010), Hall (2011) and in
Mohnen and Hall (2013). The present paper complements these reviews by providing a bibliometric
description of the how the literature addressing this particular theme has evolved in terms of contri-
bution, recognition and impact.
Our analysis paints a portrait of how research on the R&D-innovation–productivity relationships
(henceforth: RIP research) has evolved in a context of growing academic interest in each of these
three themes, and an explosion of interest in innovation in particular. We document how RIP
impact is concentrated to the discipline of Economics but also how, over time, impact disseminates
across scientificfields and extends to new communities. In particular, we detect a shift in the years
around the turn of the century where RIP research received wider attention, riding on the wave of
increased interest in the theme of innovation across the social sciences.
The present study contributes to the growing literature which attempts to write the history of
specific academic developments utilising bibliometric tools. Studying the impact and diffusion of
research through systematic cross-citation analysis techniques, such studies of the social sciences
have primarily sought to document and describe the development of interdisciplinary themes and
emerging fields of studies such as evolutionary economics (Dachs et al. 2001; Meyer 2001), innovation
(Martin 2012; Shafique 2013) and entrepreneurship (Cornelius, Landström, and Persson 2006). Fer-
reira, Pinto, and Serra (2014) apply similar techniques to a more narrowly defined object of study,
as they describe the impact of transaction cost theory on international business research. As far as
we are aware, however, this study is the first to apply bibliometric analysis to study the treatment
© 2016 Informa UK Limited, trading as Taylor & Francis Group
CONTACT Anders Broström andbr@kth.se
ECONOMICS OF INNOVATION AND NEW TECHNOLOGY, 2017
VOL. 26, NOS. 1–2, 6–20
http://dx.doi.org/10.1080/10438599.2016.1202519
Downloaded by [Kungliga Tekniska Hogskola] at 09:59 06 November 2017
of a particular research issue in academic literature. Our methodological approach is also novel in that
we demonstrate how a research approach where bibliometric analysis is combined with direct con-
frontation with key academic texts allows us to provide a broader view of the impact and diffusion of
research than a traditional literature review, but also allows a more in-depth discussion than a tra-
ditional, strictly quantitative, bibliometric analysis.
This paper proceeds as follows: Section 2 defines the research domain which is the subject of
study and discusses the RIP domain from the perspective of leading accounts of the processes of
scientific activity. In Section 3, the bibliometric methodology of the paper is accounted for. Section
4 outlines the wider development of economic literatures related to the RIP domain. Sections 5–7
proceed to delineate the impact of RIP research, through the identification of core publications
within the domain and of citations to these publications. Section 8 concludes.
2. RIP: the pursuit of a research question
The identification of prioritised research questions is a central aspect of scientific activity. In Kuhn’s
(1970) account of the history of science, each ‘paradigm’is associated with shared norms and
ideas which suggest a set of puzzles for scientists to solve and a set of tools by which to solve
these puzzles. Trying to improve the precision of solutions to these puzzles through advances in
methodology and theoretical refinement –activities which Kuhn refers to as ‘normal science’– is
what engages most scientists throughout their careers.
The object of this study –the RIP domain of scientific study on the research-innovation–pro-
ductivity relationships –can be defined through the following central research question: What are
the elasticities of firm-level productivity to the firms’R&D investment, channelled through inno-
vation? The demarcation of the domain as encompassing these three research themes is inspired
by the seminal CDM paper (Crepon, Duguet, and Mairesse 1998), which established the methodologi-
cal imperative to study the three RIP themes in conjunction in order to obtain unbiased estimates of
the elasticities. It should be noted that few if any scientific papers published within the RIP domain
are restricted in scope and aim to the issue of estimating elasticities. The CDM paper, for example,
addresses broader theoretical as well as econometric issues regarding the study of R&D activities,
innovation and firm-level productivity. RIP domain papers are nonetheless identified here through
their direct engagement with the scientific issue of determining the relationships between at least
two of these three key themes with individual firms as the level of study.
1
Studies on a wider set
of capabilities of innovation which do not directly address R&D investments are not considered to
lie in this domain. Studies on marketing or organisational innovation are furthermore not included,
unless they also study technological product or process innovation. While productivity can be
measured in levels or in terms of change over time, studies which solely address other dimensions
of firm performance such as market value are furthermore not considered to belong to the RIP
domain.
The RIP research domain is embedded in the tradition of production economics and in the field of
the economics of innovation. In the terminology of Lakatos (1970), both of these areas of study can be
classified as theories in the ‘protective belt’surrounding the hard core of the scientific research pro-
gramme of neo-classical economics (or, as suggested by Heijdra and Lowenberg (1986), as sub-dis-
ciplinary demi-cores). As such, there appears to be widespread acceptance of the general relevance
of RIP research among economists. Martin (2012), for example, writes about the problem of under-
standing the returns to R&D as ‘a central building block’for studies on innovation. The publication
of RIP contributions such as Mansfield (1980) and Griliches (1994) in the leading journal American
Economic Review also signals that the problem has been considered broadly relevant. It is also
noteworthy that even in leading early contributions to the literature by authors such as Mansfield,
Griliches, Hall and Mairesse (see references in Appendix 1), motivations for the relevance of the
RIP research problem(s) are –where at all touched upon –limited to a general acknowledgement
of significant scholarly interest in the issue. This is indicative of RIP research as being strongly
ECONOMICS OF INNOVATION AND NEW TECHNOLOGY 7
Downloaded by [Kungliga Tekniska Hogskola] at 09:59 06 November 2017
embedded in the research program which has held a dominating position within the economic
sciences for several decades.
As an integral part of a successful research program and a line of research with a history span-
ning several decades, the RIP domain constitutes an interesting example of scientificactivityinthe
social sciences. The research problem of determining elasticities is also somewhat unusual in char-
acter for the social sciences, in that it concerns a problem of assessing a magnitude. Even in the
field of Economics, theories and their empirical evaluations are typically more concerned with
the direction of causal influence than with the level of impact (Wade 2007). The interest in estab-
lishing magnitudes have seemingly spurred scholars to continuously re-visit the central research
questions of the RIP domain equipped with new empirical data and updated estimation method-
ology. In what follows, we will apply bibliometric methods to analyse the imprint of RIP research on
the scientific literature.
3. Bibliometric methodology
We utilise bibliometric techniques to study the context within which RIP research has evolved and the
impact of RIP research on the wider scientific literature. The domain as such is for this purpose rep-
resented by a set of ‘core’papers, which are identified as the most frequently cited contributions in
the domain. Citations to these papers were used to track the impact of RIP research on the wider
scientific literature.
The Web of Science (WoS) database was chosen as the main material for analysis, as it has good
global coverage of scientific publications, relatively strict quality threshold criteria for inclusion of
scientific journals and a widely recognised categorisation of scientificfields. The searches in WoS
were restricted to 1990–2012. When using keyword searches, the starting year 1992 was selected
since abstracts were introduced in the database this year.
2
Beyond abstracts, searches for keywords
also included the title of individual papers.
The citation rates where field normalised, that is, the mean number of citations to a paper were
divided by mean number of citations for all papers in the same field in the same year. Here we
used the 251 Web of Science Categories for normalisation. Separate field norms are used for articles
and reviews. This means that a mean field normalised citation rate of all papers in a field is 1 and for
example, a citation rate of 1.50 means a citation rate 50% higher than world average.
Terms are extracted and mapped using the VOSviewer software (http://www.vosviewer.com).
4. A wave of innovation research
Our analysis of RIP research starts in a bibliometric analysis of how the interest in the three core
themes of R&D, innovation and productivity has shifted over time. Figure 1 demonstrates the
growth of innovation and R&D as research themes in bibliometric terms. Panel B shows how,
between 1990 and 2012, the number of papers addressing innovation or R&D has increased 15-
fold. While this development certainly is impressive by any measure used, it has to be related to
the general volume growth of research output during this period. As worldwide spending on aca-
demic research has increased, publishing patterns have changed towards increased emphasis on
publishing in international rather than national journals and in papers rather than monographs,
the number of journals and papers have risen throughout the social sciences. Parallel to this devel-
opment, the number of journals indexed in the WoS database in general and the number of papers in
within Social Sciences in general as well as in the sub-field Business and Economics has increased
substantially (see panel A). This growth, however, is dwarfed by that of scientific attention to inno-
vation and R&D, which grew eight times as much as the Business & Economics field in total (see
panel B). The proportion of papers addressing innovation or R&D in WoS has increased, from less
than 2% of B&E in 1990 to almost 13% in 2012 (Figure 1, panel C).
8A. BROSTRÖM AND S. KARLSSON
Downloaded by [Kungliga Tekniska Hogskola] at 09:59 06 November 2017
In parallel with this increase there has been a global shift in where the papers are produced.
3
In the
1990s USA and Canada produced more than half (53%) of all papers in this field, followed by Europe
(with 36%) while the rest of the world contributed 11% of the papers. Currently (2010–2012), one
third of the papers are produced in countries outside North America and Europe. This relative
increase comes at the expense of North American dominance (22%) while the European proportion
increased slightly (to 45%).
In Figure 2, the development of the three main keywords of the R&D-innovation-productivity
research theme is shown for each term separately.
4
While all three keywords show a growing
trend over the studied period, suggesting an increasing attention paid to all three concepts by scho-
lars in the Business & Economics field, there are marked differences in terms of growth. In 1990, about
2% of all papers in the B&E field addressed issues of innovation and of productivity, respectively. By
2012, the interest in productivity has steadily grown so that over 3% of all B&E papers list this term as
a keyword. The term innovation, however, was by then provided as keyword on about 8% of all B&E
papers. All three keywords show the same shift in the geography of the paper production from USA
to other parts of the world, Europe maintaining a fairly constant fraction over time.
Behind these growth patterns, a difference in how broadly the respective themes have been
picked up across disciplines in the Business & Economics field can be surmised. The study of pro-
ductivity remains strongly concentrated to the sphere of Economics. The interest for innovation,
on the other hand, has penetrated the entire business studies community, surfacing in studies of
economic geography, business history, industrial dynamics and several other sub-disciplines. As a
widely recognised research theme, the study of innovation has developed along tracks where the
connection to the other two concepts is less often in focus. Prominent examples are the focus on
analysis of the systemic nature of innovation (Freeman 1987) and the strand of literature which
explores antecedents of innovation other than that of formal R&D (Pavitt 1998). While many
Figure 1. The general development of Social Sciences, the Business and Economics field and innovation- R&D-literature in WoS
1990–2012. Legend: The inserted pie charts show the geographic distribution of the papers among Europe (light grey), North
America (medium grey) and other parts of the world (dark grey) in the two periods 1992–1999 and 2010–2012.
Figure 2. Frequencies of papers featuring three main keywords as a percentage of all papers within the Business & Economics field
in WoS 1990–2012.
ECONOMICS OF INNOVATION AND NEW TECHNOLOGY 9
Downloaded by [Kungliga Tekniska Hogskola] at 09:59 06 November 2017
influential studies in both of these strands of literature addressed innovation at other levels of analy-
sis and through other empirical methodology than that associated with the RIP domain, the growth
of general interest in innovation as a research theme has certainly had an impact on scholars’interest
in conducting research in the RIP domain and paved the way for increased impact of the domain
across the social sciences. In what follows, we analyse the volume and nature of this impact.
5. A bibliometric representation of the RIP domain
The bibliometric analysis of the RIP domain is based on the identification of a set of key papers within
the domain. For this purpose, the WoS data source was judged to be too restrictive in terms of
content, as several papers widely acknowledged as key contributions in RIP research have been pub-
lished as working papers or in journals hitherto not included in WoS. Therefore the databases SCOPUS
and Google scholar were used to identify key RIP papers. SCOPUS has the advantage of including a
larger journal set than WoS, while also including many books and conference papers. Google scholar
is by far the broadest of the three sources considered here, listing individual chapters in books and
working papers. Searches on well-known RIP contributions also suggested to us that Google scholar
complemented SCOPUS limitations on records from before 1996 (c.f. Jacso 2005).
In view of these database characteristics, two complementary search methods were use. First, we
searched SCOPUS for all papers with more than 100 listed citations by February 2015 featuring at
least two of the terms ‘R&D’,‘innovation’and ‘productivity’in their title, abstract or among listed key-
words. In complementary searches, ‘R&D’was substituted by ‘research’, innovation by ‘patent’,‘tech-
nical change’and ‘technological change’and productivity by ‘performance’. In total, the 19 searches
on combinations of the above terms restricted to the three SCOPUS subject fields ‘Economics, Econo-
metrics and Finance’,‘Social science’and ‘Engineering’, rendered 79 papers fulfilling the citation
threshold criterion. Out of these, 15 were found to belong to the RIP domain.
Second, we searched Google scholar for all papers on the reference lists of three RIP reviews (Mair-
esse and Sassenou 1991; Hall 2011; Mohnen and Hall 2013) as well as the oldest regular RIP paper
identified in the previous stage (Griliches 1979). Papers with more than 500 citations in the
Google scholar database by February 2015 were selected, whereof five were identified as RIP
papers not identified through the first search method. Repeating this procedure on the references
listed in the RIP papers identified so far, another set of eight publications with more than 500
Google scholar citations was identified.
In total, searches thus rendered 28 publications. 8 of these are published before 1990, 9 are pub-
lished in the 1990s and a further 11 in the first decade of the twenty-first century. In 12 instances, a
working paper version (often separated in time) with an identical name was identified. These refer-
ences were also added to our list. The complete list of identified papers is provided in Appendix 1.
6. Impact of RIP research: journals, authors, disciplines
To map the influence of RDIP research, we next identify a set of 3274 papers in WoS which cite at least
one of the 28 core papers. The annual number of citing papers increases strongly and continuously
over the studied period, from an average of 17 papers per year in 1990–1994 to 390 papers per year in
the period 2008–2012. This increase strongly outperforms the general growth of economic literature
included in the WoS which, as shown in Figure 1, grew about 150% in volume over the same period.
To some extent, an increase in citations to the full set of 28 papers over time is to be expected, as the
set –which was not designed with the primary purpose of studying temporal patterns –consists of
papers published in the time-span 1979–2004. Papers naturally do not accumulate citations until they
are published. Nonetheless, the growth of citation to the core RIP publications indicates a strong and
persistent increase in attention to the domain during the studied period.
The set of 3274 citing papers have 4916 unique authors and are published in 745 different jour-
nals. Table 1 shows the names of the ten most prolific of these authors. Table 2 lists the 20 most
10 A. BROSTRÖM AND S. KARLSSON
Downloaded by [Kungliga Tekniska Hogskola] at 09:59 06 November 2017
frequently occurring journals, accounting for about a third of all publications. Publications are fairly
broadly distributed across journals, but with a significant concentration to the journal Research
Policy, which alone accounts for 9% of the publications.
5
Table 3 shows how these papers are distributed across WoS subject classifications (Rafols and Leydes-
dorff 2009). We find that a lion’s share of all papers drawing on RIP research have been published in
journals classified as belonging to the field of Economics.
6
Furthermore, significant shares of the
papers are published in journals which are classified as belonging to other economic disciplines (Man-
agement, Business, Planning & Development). Temporal analysis (not showed in the table) identifies a
shift in patterns occurring around the year 2000. After this year, the share of citing papers published
in journals classified as Economics and as ‘Social Sciences, Mathematical Methods’fall back. The share
of papers originating from Management and Business publications increase correspondingly.
From the right-hand column of Table 3 it is clear that the citation impact of the paper set used for
this analysis is highly cited, with publications in the five most frequent fields cited between 45% and
66% above world field average. The papers citing the core-set of papers are thus themselves making
a significant impact in a range of subject field categories.
While a vast majority of impact is to be found within economic fields, we note that citations are
found across the map, received from 94 out of 222 categories. We note that the average citation level
is significantly higher for papers published in journals which, judging by their classification, are
oriented towards the methodological aspects of RIP research (‘Mathematics, Interdisciplinary Appli-
cations’,‘Social Sciences, Mathematical Methods’and ‘Statistics & Probability’) than for papers pub-
lished in journals which are more explicitly embedded within the social sciences.
Table 1. Top 10 authors of papers citing core RIP contributions.
Author Number of papers
Yang, C.H. 24
Czarnitzki, D. 19
Hall, B.H. 18
Lerner, J. 15
Van Reenen, J. 16
Vivarelli, M. 15
Roper, S. 14
Tsai, K.H. 14
Gamberdella, A. 13
Love, J.H. 13
Table 2. Top 20 journals publishing papers which cite core RIP contributions.
Journal Number of papers
Research Policy 280
Strategic Management Journal 67
Industrial and Corporate Change 62
Small Business Economics 59
Scientometrics 57
Applied Economics 56
Technovation 49
International Journal of Industrial Organization 44
International Journal of Technology Management 42
Management Science 42
Technological Forecasting and Social Change 42
Review of Economics and Statistics 40
American Economic Review 37
Industry and Innovation 31
Journal of Evolutionary Economics 30
Journal of Technology Transfer 30
Organization Science 30
Journal of International Business Studies 28
R&D Management 28
Subtotal number of papers in 20 most frequently occurring journals 1054
Total number of papers 3274
ECONOMICS OF INNOVATION AND NEW TECHNOLOGY 11
Downloaded by [Kungliga Tekniska Hogskola] at 09:59 06 November 2017
In Appendix 2, the different WoS subject field categories listed in Table 3 are visualised in relation
to each other using Rafols and Leydesdorff (2009)’s World Map of Science. This illustration demon-
strates that that some of the fields listed in Table 3 above are located at a fair distance from the
field of Economics. This implies that a small but significant share of all citations to the 28 core RIP
papers come from scientificfields where linkages to the field of Economics through co-citations
are unusual, as judged from the entirety of papers included in the WoS database.
We suggest that these citation pattern demonstrates that the core of the RIP research program has
been pushing the methodological frontiers, attracting attention and creating impact in research com-
munities where development of quantitative analysis of economic activity is of greater interest than
the issue of estimating the elasticities of the R&D-innovation–productivity relationships.
7. Impact of RIP research: thematic clustering
We next characterise the set of citing papers by means of analysing which terms are most frequently
occurring in titles and abstracts. Using a threshold of minimum 30 occurrences for terms to be
included, 353 terms where identified using the VOSviewer software (Van Eck and Waltman 2011).
As expected, the terms used to identify the gross list of core RIP papers also feature prominently. ‘Pro-
ductivity’appears in 21% of all papers, ‘R&D’in 36%, ‘patent’in 29% and ‘innovation’in 38%. The
occurrence of these terms over time is reasonably stable, with one notable exception. The term ‘inno-
vation’features in 15% of the 300 papers in the set which are published before 2001, and in a whop-
ping 47% of the 1984 papers published after 2007. This finding supports the claim (see Section 4
above) that the broadened interest in RIP research has been carried by a surge of interest in the
notion of innovation during the first decade of the twenty-first century.
The extraction of terms from the title an abstract of papers can also be used to establish relaton-
ships between papers in the set of citing papers. Figure 3 shows a term-density map created using
the VOSviewer software. Terms occurring in the same document are placed close to each other and
frequently mentioned terms are displayed in a larger font size than more infrequently occuring terms.
Only terms which effectively identifies clusters of papers are shown on the map.
Afirst interesting observation is that the term ‘innovation’is not found to define a demarcated
cluster of publications, but occurs in combination with other terms (e.g. ‘innovation activity’and
‘innovation performance’) scattered across the lower left quadrant of the map. This likely reflects
that the concept of innovation has been widely adopted and occurs in such a broad set of publi-
cations that it does not meaningfully define specific research niches in bibliometric analysis.
The term-density map shows three particular concentration points, which define separable cluster
of publications. On the upper left side of Figure 3, the terms ‘productivity’and ‘R&D investment’
feature together. Papers who are themselves in the RIP domain constitute an important part of this
Table 3. Top 15 fields from which core RIP contributions are cited.
Field Map code Number of papers Mean field normalised citation rate
Economics 1 1638 1.45
Management 2 1230 1.59
Business 3 681 1.44
Planning & Development 4 422 1.65
Operations Research & Management Science 5 215 1.66
Engineering, Industrial 6 204 1.13
Business, Finance 7 160 1.58
Social Sciences, Mathematical Methods 8 117 2.07
Environmental Studies 9 117 0.94
Information Science & Library Science 10 103 1.11
Computer Science, Interdisciplinary Applications 11 91 0.91
Law 12 89 2.25
Geography 13 68 1.14
Engineering, Multidisciplinary 14 54 0.41
International Relations 15 51 0.98
12 A. BROSTRÖM AND S. KARLSSON
Downloaded by [Kungliga Tekniska Hogskola] at 09:59 06 November 2017
cluster, together with highly related studies on industry-level elasticities between research,
innovation and productivity. The central upper cluster is focused on the terms ‘growth’and ‘rate’.
This cluster is closely tied to the first; the composite term ‘productivity growth’features in the title or
abstract of 9% of the citing papers. That this cluster is identified as a separate group by the algorithm
reflects how RIP core papers have been cited in works discussing the role of R&D and innovation for
macroeconomic development. The composite terms ‘economic growth’occurs in as many as 8% of all
citing papers.
Finally, in the lower right-hand side of the figure we find a clearly separated cluster of papers
linked together by the occurrence of terms related to patenting. Some of the papers in this cluster
are in the RIP domain, where the count of patents or patent applications has become the most
common measure of innovation output. But interestingly, we also find a broad set of papers relating
to RIP research though citation of at least one of the ‘core’paper in connection to research analysing
the patent system as such and the immaterial property rights (IPR) strategy of firms.
8. Impact of RIP research: citation function analysis
Classification of publications by means of bibliometric tools and techniques such as those applied above
provide good overview and orientation of a set of scientific papers, but clearly also leave many questions
unanswered. In order to provide a more in-depth analysis of the impact of RIP research, we analyse of in
what context the set of 28 core RIP is cited by other papers. For this purpose, we delimit the set of 3274
Figure 3. Cluster analysis of papers citing the set of 28 core RIP publications, by terms featuring in abstracts and titles.
ECONOMICS OF INNOVATION AND NEW TECHNOLOGY 13
Downloaded by [Kungliga Tekniska Hogskola] at 09:59 06 November 2017
citing papers used for the analysis in the previous two sections to the sub-set those papers which also
themselves have received at least 100 citations from WoS papers. Hundred and twenty-six citing papers,
and 184 unique citing–cited paper pairs are identified in this manner. Several citing papers contain mul-
tiple references to the same cited paper. In parallel to the bibliometric analysis of the previous two sec-
tions, we do however in what follows refer to each of the 184 pairs as a citation.
Afirst finding, after manually reading and categorizing these papers, is that only 12 papers (9%)
are themselves within the RIP domain. This finding complements the disciplinary and thematic ana-
lyses of the influence of RIP research, as presented above, in providing a picture of RIP research as
highly relevant outside the RIP domain itself.
We also carefully study all citations to the 28 core RIP papers in the 126 citing papers in order to
analyse the function of the cited paper in reference to the citing paper. In the literature on the prac-
tice of scientific citation, several categorization schemes for such analysis have been suggested and
applied. Peritz (1983), for example, presents eight functions in which citations are used in empirically
oriented social sciences.
7
For the purpose of characterizing the impact of RIP research by citation
function, we apply a simplified scheme of analysis structured around the following two questions:
Question 1: Does the citing paper build on results and arguments from the cited paper to establish
a position about a real-world phenomenon relevant to the present study?
From a position of what in the philosophy of science is referred to as realism, such citations may be
referred to as filling an evidential function. The cited paper is referred to as having advanced a specific
idea, to have shown/suggested a causal relationship (i.e. establishing which control variables to use in
econometric modelling exercises) and more generally as presenting arguments or findings that are
directly relevant for the scientific work of the citing paper. The citing paper references the cited paper
in relation to an affirmative statement of an ontological nature.
For a reference where the question above was assessed negatively, a second question is posed.
8
Question 2: Does the citing paper build on the cited paper in order to establish how something
can or should be researched?
In the context of papers citing RIP research, a vast majority of such citations are related to the
empirical issues such as the choice of estimators, modelling techniques, and data. Citations with
this type of methodological function are also used in problematizing (e.g. in reference to measure-
ment problems), conceptualising (e.g. in reference to the relationship between theoretical constructs
and empirical measures), or explanatory statements (e.g. in discussing alternative theoretical mech-
anisms behind an observed empirical regularity).
Following Moravcsik and Murugesan (1975), we refer to citations which fill neither of these two
functions as perfunctory. Such references are for example used to acknowledge the general existence
of a paper without further explaining how or to what extent the citing paper relates to the cited
paper. A handful of perfunctory references are also provided to papers cited as having neglected
to study a specific aspect of a phenomenon.
We are interested in mapping how each of the citing papers relates to the cited source paper. Classi-
fication of individual references as per above are therefore aggregated to the level of citing papers. Table
4shows how citations are distributed between the three categories of citation functions listed above.
In light of the scarcity of available studies on citation functions within the social sciences against
which the results of Table 4 may be compared, a certain degree of caution must be applied in
Table 4. Function of citations to core RIP papers.
Percentage of all citing papers
Share of papers only containing citation(s) used to establish
a position about a real-world phenomenon
47 %
Share of papers only containing citation(s) used in order to establish
how something can or should be researched
27 %
Share of papers containing at least one citation of each
kind identified in the paper
10 %
Share of papers containing perfunctory citation(s) only 17 %
14 A. BROSTRÖM AND S. KARLSSON
Downloaded by [Kungliga Tekniska Hogskola] at 09:59 06 November 2017
interpreting the results. Nonetheless, we are prepared to argue that the RIP domain is set apart from
empirically oriented research in general by the finding that 37%
9
of papers citing RIP research do so
in the context of motivating choices for research design and methodology. We note that at 17%, the
percentage of perfunctory references is relatively low, at least in comparison to previous studies of
the physical sciences where the corresponding figure has been found to approach 50%. The lack
of available studies of citation functions in the social sciences makes it difficult to draw strong con-
clusions about whether this result is a particularity of the RIP domain, or whether this number is repre-
sentative for the economic field. We may, however, conclude that a large majority of the citational
linkages identified in Sections 5 and 6 are indeed to be understood as an indication of intellectual
debt to RIP papers, not merely as an acknowledgement of their existence.
In order to obtain some points of reference against which the results on citations to core RIP
papers may be compared, we analyse further sets of citations. In particular, we are interested in
obtaining benchmarks on methodological and perfunctory citation practises. Table 5 presents
results from three such exercises. To enable direct comparison with the results from Table 4,we
include the summation of rows 2 and 3 in Table 4 in the upper left-hand cell.
Afirst benchmark is provided by all papers published in three recent issues (Issue 7 (2014), Issue 8
(2014), and Issue 1–2 (2015)) of Economics of Innovation and New Technology. The citations of these
papers are analysed in a parallel fashion to the analysis presented in Table 4. In this set, one out of
three papers was categorised as having at least one reference filling a methodological function. In
one paper out of five, we found perfunctory citations only. However, a non-trivial share of such cita-
tions seems to stem from papers strictly focused on aspects of how something may be researched
(statistical methodology and issues of theoretical economic modelling).
Complementing this first benchmarking exercise, we undertake a second round of analysis
focused on highly cited empirical papers. From each of the journal issues where the first four (in
alphabetic order) papers on our list of 28 core papers were published, we identify all papers featuring
empirical microeconomic analysis. Among those, we selected one paper from each issue that most
closely matches the RIP paper published in the same issue in terms of citation counts. The resulting
set of papers is listed in Appendix 3. For this set, almost a third of citations are perfunctory, and less
than 1 citation in 10 is classified as having a methodological function, that is, to be used only in order
to establish how something can or should be researched.
In further investigation, we conduct a third wave of citation function analysis, this time with a par-
ticular core RIP paper in focus. The CDM article (Crepón, Duguet, and Mairesse 1998) is widely
acknowledged as a methodological milestone paper both within the RIP domain per se and in the
wider fields of production economics and the economics of innovation (see Lööf, Mohnen, and Mair-
esse, forthcoming).
10
We find that a full three out of five of all citations to either one of the two ver-
sions of this paper have a methodological function. Two indications of the CDM paper’s central
importance for the work presented in many citing papers is that we find many instances of extensive
referencing among the citing papers and that in well beyond one out of four of all papers, we find
citations filling both types of constructive functions identified in the first two rows of Table 4.
The comparisons of Table 5 suggest that in comparison to related research streams, the frequency
of perfunctory citations is relatively low and the frequency of methodological citations fairly high for
core RIP contributions. We interpret this as an indication that RIP research has made significant meth-
odological contributions.
Table 5. Function of citations to core RIP papers.
Core RIP
papers (%)
Comparison 1:
EINT (%)
Comparison 2:
Empirical papers (%)
Comparison 3:
CDM (%)
Share of papers containing citation(s) used in order to
establish how something can or should be researched
37 32 9 61
Share of papers containing perfunctory citation(s) only 17 21 32 10
ECONOMICS OF INNOVATION AND NEW TECHNOLOGY 15
Downloaded by [Kungliga Tekniska Hogskola] at 09:59 06 November 2017
9. Conclusions
Research on the relationship between a firm’s investment in R&D, its success in introducing inno-
vations and it’s performance in terms of productivity have a long history. Utilising a novel combi-
nation of bibliometric analysis and direct engagement with academic texts, this paper has
accounted for the impact that this area of study (‘the RIP domain’) has had on the scientific literature
over three decades. We show that the share of papers within the economic sciences with the term
‘innovation’featured in the title or abstract increased fourfold between 1990 and 2012, with a
clear trend shift occurring around the turn of the millennium. In this intellectual environment, interest
in the RIP domain was increased and broadened across scientificfields.
We argue that the strong embeddedness of the domain in the tradition of neo-classical pro-
duction economics and the (neo-classical) economics of innovation provided a basic academic legiti-
macy for RIP research, but the theory-independent nature of the core RIP research question may have
increased the wider accessibility of research results originating from the domain. The interest in RIP
research was probably also positively affected by a simultaneous increase in interest in the causes of
inter-firm productivity across the entire field of Economics (Syverson 2011) and fuelled by signifi-
cantly improved availability of firm-level data of high quality in many countries, e.g. though the diffu-
sion of the Community Innovation Surveys (Arundel 2008).
Indicative of this broadened interest in the research-innovation–productivity relationship, we find
indications from keyword analysis (see Section 7) that macroeconomic studies and industry-level
analysis has drawn rather extensively on RIP research. We also find that 58% of citations to RIP
research are made in a context where the author(s) build on empirical results or theoretical ideas
about economic realities presented from the cited paper. Only a limited fraction of citations are
made in general acknowledgement of the field and its central papers.
RIP research has moved forward through constant refinement of estimation methodology, careful
re-interpretation of the conceptual linkages between theoretical constructs and empirical measure-
ment and though the collection of richer data sets. Thereby, research within the domain has made a
number of important contributions to the wider scientific literature. We show that about one out of
three citations to RIP papers are made in reference to methodological issues. Cluster analysis depicts
studies analysing the institutions of IPR and the IPR strategies of firms as constituting an important
group among all papers citing RIP research. We also find that papers from journals which, judging by
their classification, are oriented towards the methodological aspects of RIP research is significantly
higher than the average impact of citing papers. Together, these findings suggest that the methodo-
logical contributions of RIP research have broadened the interest in the domain.
The main contribution of the present paper is to provide a comprehensive overview of academic
research into the research-innovation–productivity relationship. We do, however, also hope to have
contributed to methodological development in the field of science studies. In particular, we have
shown that specific research problems and research domains may be subject to analysis of a kind
previously applied to wider fields or theories, and that such study can provide important insights
into the workings of the international scientific system. We also hope to have illustrated that a com-
bination of bibliometric techniques and direct study of academic texts is a valuable complement both
to the traditional research review and the traditional bibliometric study.
Notes
1. In 1970s and 1980s RIP work, where the research group formed around Zvi Griliches at NBER played a leading role,
the central issue was that of understanding and assessing the returns to R&D. The theme of innovation was, as
pointed out later, advanced at a later stage and firmly incorporated into the RIP domain e.g. through the CDM study.
2. In 1992, 64% of all papers in the field ‘Business and Economics’in Web of Science had an abstract, this value
increased successively to about 96% in 2012. Before 1992, keyword searches resulted in very few found records.
3. Changes in global paper production is expressed as the proportion of fractionalized papers coming from Europe,
North America (US or Canada) or other parts of the world. The fractionalization means that when there are authors
16 A. BROSTRÖM AND S. KARLSSON
Downloaded by [Kungliga Tekniska Hogskola] at 09:59 06 November 2017
from several countries on a paper, each region is credited a fraction of each paper in proportion to the number of
addresses to each region.
4. Notably, only a small fraction of the WoS papers identified above include all of the three keywords in their abstract
and/or title. For the publication year 2012, 130 such articles can be identified within the Business & Economics
field, with an additional 100 articles in the wider WoS database.
5. Note that the EINT journal does not appear in Table 2 since it has not yet been included in the WoS database.
6. In the Web of Science database, journals may be classified in more than one subject field category.
7. These are (1) Setting the stage; (2) Background information; (3) Methodological; (4) Comparative; (5) Argumental,
speculative, hypothetical; (6) Documentary; (7) Historical; (8) Casual.
8. Ontological questions about economic realities obviously are central for considerations how economic phenom-
ena should be researched. While not figuring in a methodological context, all citations of an evidential nature can
therefore be considered to be relevant for questions about how an issue can or should be researched.
9. This is the sum of rows two and three in Table 4.
10. Citing papers are as before identified through WoS. In order to obtain a larger set of analysable citing papers, we
lower the threshold inclusion criteria from 100 WoS citations to 30 for this analysis. Forty-three paper citing
Crepón, Duguet, and Mairesse (1998) are identified in this manner.
Disclosure statement
No potential conflict of interest was reported by the authors.
ORCiD
Anders Broström http://orcid.org/0000-0003-0820-2769
Staffan Karlsson http://orcid.org/0000-0002-5739-5213
References
Arundel, A. 2008.“Innovation Surveys and Policy: Lessons from the CIS.”In Innovation policy in Europe: Measurement and
Strategy, edited by C. Nauwelaers and R. Wintjes, 3–28. Cheltenham: Edward Elgar Publishing.
Cornelius, B., H. Landström, and O. Persson. 2006.“Entrepreneurial Studies: The Dynamic Research Front of a Developing
Social Science.”Entrepreneurship Theory and Practise 30 (3): 375–398.
Crepon, B., E. Duguet, and J. Mairesse. 1998.“Research, Innovation and Productivity: An Econometric Analysis at the Firm
Level.”Economics of Innovation and New Technology 7 (2): 115–158.
Dachs, B., T. Roediger-Schluga, C. Widhalm, and A. Zartl. 2001.“Mapping Evolutionary Economics: A Bibliometric Analysis.”
Paper presented at the EMAEE conference, Vienna, Austria, September 13–15.
Ferreira, M. P., C. F. Pinto, and F. R. Serra. 2014.“The Transaction Costs Theory in International Business Research: A
Bibliometric Study Over Three Decades.”Scientometrics 98 (3): 1899–1922.
Freeman, C. 1987.Technology Policy and Economic Performance: Lessons from Japan. London: Frances Pinter.
Griliches, Z. 1979.“Issues in Assessing the Contribution of Research and Development to Productivity Growth.”Bell
Journal of Economics 10 (1): 92–116.
Griliches, Z. 1994.“Productivity, R&D, and the Data Constraint.”The American Economic Review 84 (1): 1–23.
Hall, B. H. 2011.Innovation, and Productivity. NBER Working Paper No. 17178. Cambridge, MA: NBER.
Hall, B. H., J. Mairesse, and P. Mohnen. 2010.“Measuring the Returns to R&D.”In Handbook of the Economics of Innovation.
Vol. II, edited by B. H. Hall and N. Rosenberg, 1033–1082. Burlington, MA: Academic Press.
Heijdra, B. J., and A. D. Lowenberg. 1986.“Duhem-Quine, Lakatos and Research Programmes in Economics.”The Journal of
Interdisciplinary Economics 1 (3): 175–187.
Jacso, P. 2005.“As We May Search –Comparison of Major Features of the Web of Science, Scopus, and Google Scholar
Citation-Based and Citation-Enhanced Databases.”Current Science 89 (9): 1537–1547.
Kuhn, T. 1970.The Structure of Scientific Revolutions. 2nd ed. Chicago, IL: University of Chicago Press.
Lakatos, I. 1970.“Falsification and the Methodology of Scientific Research Programmes.”In Criticism and the Growth of
Knowledge, edited by I. Lakatos and A. Musgrave, 91–196. Cambridge: Cambridge University Press.
Lööf, H., J. Mairesse, and P. Mohnen. Forthcoming. Introduction: CDM 20 years after. Economics of Innovation and New Technology.
Mairesse, J., and M. Sassenou. 1991.R&D Productivity: A Survey of Econometric Studies at the Firm Level. NBER Working
Paper No. 3666. Cambridge, MA: NBER.
Mansfield, E. 1980.“Basic Research andProductivity Increases in Manufacturing.”American Economic Review 70 (5): 863–873.
Martin, B. 2012.“The Evolution of Science Policy and Innovation Studies.”Research Policy 41 (7): 1219–1239.
Meyer, M. 2001.“Nelson and Winter’s Evolutionary Theory –A Citation Analysis.”SPRU, University of Sussex, Mimeo.
Mohnen, P., and B. H. Hall. 2013.“Innovation and Productivity: An Update.”Eurasian Business Review 3 (1): 47–65.
ECONOMICS OF INNOVATION AND NEW TECHNOLOGY 17
Downloaded by [Kungliga Tekniska Hogskola] at 09:59 06 November 2017
Moravcsik, M. J., and P. Murugesan. 1975. “Some Results on the Function and Quality of Citations.”Social Studies of Science
5 (1): 86–92.
Pavitt, K. 1998.“Technologies, Products and Organization in the Innovating Firm: What Adam Smith Tells Us and Joseph
Schumpeter Doesn’t.”Industrial and Corporate Change 7 (3): 433–452.
Peritz, B. C. 1983. “A Classification of Citation Roles for the Social Sciences and Related Fields.”Scientometrics 5 (5): 303–
312.
Rafols, P. A., and A. Leydesdorff. 2010.“Science Overlay Maps: A New Tool for Research Policy and Library Management.”
Journal of the American Society for Information Science & Technology 61 (9): 1871–1887.
Shafique, M. 2013.“Thinking Inside the Box? Intellectual Structure of the Knowledge Base of Innovation Research.”
Strategic Management Journal 34 (1): 62–93.
Syverson, C. 2011.“What Determines Productivity?”Journal of Economic Literature 49 (2): 326–365.
Van Eck, N. J., and L. Waltman. 2011.“Text Mining and Visualization Using VOSviewer.”ISSI Newsletter 7: 50–54.
Wade, D. W. 2007.“Popper and Lakatos in Economic Methodology.”In The Philosophy of Economics, edited by D.
Hausman, 188–204. Cambridge: Cambridge University Press.
Appendix 1: Key papers.
Acemoglu, D., and J. Linn. 2004. “Market Size in Innovation: Theory and Evidence from the Pharmaceutical Industry.”
Quarterly Journal of Economics 119 (3): 1049–1090.
Basant, R., and B. Fikkert. 1996. “The Effects of R&D, Foreign Technology Purchase, and Domestic and International Spil-
lovers on Productivity in Indian firms.”Review of Economics and Statistics 78 (2): 187–199.
Bloom, N., and J. Van Reenen. 2002. “Patents, Real Options and Firm Performance.”The Economic Journal 112 (478): C97–C116.
Bound, J., C. Cummins, Z. Griliches, B. H. Hall, and A. Jaffe. 1984. “Who Does R&D and Who Patents?”In R&D, Patents, and
Productivity, edited by Zvi Griliches. Chicago: University of Chicago Press. Also published in 1982 as NBER Working Paper
No. 908.
Cohen, W. M., and S. Klepper S. 1996. “A Reprise of Size and R & D.”The Economic Journal 106 (437): 925–951.
Crepon, B., E. Duguet, and J. Mairesse. 1998. “Research, Innovation and Productivity: An Econometric Analysis at the Firm
Level.”Economics of Innovation and New Technology 7 (2): 115–158. Also published in 1998 as NBER Working Paper No. 6696.
Branstetter, L. G. 2001. “Are Knowledge Spillovers International or Intranational in Scope? Microeconometric Evidence
from the U.S. and Japan.”Journal of International Economics 53 (1): 53–79.
Goto, A., and K. Suzuki. 1989. “R&D Capital, Rate of Return on R&D Investment and Spillover of R&D in Japanese Manu-
facturing.”Review of Economics and Statistics 71 (4): 555–564.
Griffith, R., E. Huergo, J. Mairesse, and B. Peters. 2006. “Innovation and Productivity Across Four European Countries.”
Oxford Review of Economic Policy 22 (4): 483–498.
Griliches, Z. 1979. “Issues in Assessing the Contribution of Research and Development to Productivity Growth.”Bell
Journal of Economics 10 (1): 92–116. Also published in 1998 as chapter in Griliches, Z., R&D and Productivity: The Econo-
metric Evidence, University of Chicago Press.
Griliches, Z. 1986. “Productivity, R&D, and Basic Research at the Firm Level in the 1970s.”American Economic Review 76 (1):
141–154. Also published in 1985 as NBER Working Paper No. 1547.
Griliches, Z. 1998. R&D and Productivity: The Econometric Evidence. University of Chicago Press.
Griliches, Z. 1994. “Productivity, R&D, and the Data Constraint.”American Economic Review 84 (1): 1–23. Also published in
1998 in Griliches, Z., R&D and Productivity: The Econometric Evidence, 347–374. University of Chicago Press, NBER Books.
Griliches, Z., and J. Mairesse. 1984. “Productivity and R and D at the Firm Level.”In R&D, Patents and Productivity, edited by
Z. Griliches, 339–374. University of Chicago. Also published in 1981 as NBER Working Paper No. 826.
Griliches, Z., and J. Mairesse. 1998. “Production Functions: the Search for Identification.”In Econometrics and Economic
Theory in the 20
th
Century, edited by S. Strom. Cambridge: Cambridge University Press. Also published in 1995 as
NBER Working Paper No. 5067.
Hall, B. H., Z. Griliches, and J. A. Hausman. 1986. “Patents and R&D: Is There a Lag?”International Economic Review 27 (2):
265–283. Also published in 1984 as NBER Working Paper No. 1454.
18 A. BROSTRÖM AND S. KARLSSON
Downloaded by [Kungliga Tekniska Hogskola] at 09:59 06 November 2017
Hall, B. H., and J. Mairesse. 1995. “Exploring the Relationship Between R&D and Productivity in French Manufacturing
Firms.”Journal of Econometrics 65 (1): 263–293.
Hall, B. H., and R. H. Ziedonis. 2001. “The Patent Paradox Revisited: An Empirical Study of Patenting in the U.S. Semicon-
ductor Industry, 1979–1995.”RAND Journal of Economics 32 (1): 101–128.
Henderson, R., and I. Cockburn. 1996. “Scale, Scope, and Spillovers: The Determinants of Research Productivity in Drug
Discovery.”RAND Journal of Economics 27 (1): 32–59. Also published in 1993 as NBER Working Paper No. 4466.
Hu, A. G. Z., G. H. Jefferson, and Q. Jinchang. 2005. “R&D and Technology Transfer: Firm-Level Evidence from Chinese
Industry.”Review of Economics and Statistics 87 (4): 780–786.
Klette, T. J., and S. Kortum. 2004. “Innovating Firms and Aggregate Innovation.”Journal of Political Economy 112, 86–1018.
Also published in 2002 as NBER Working Paper No. 8819.
Kortum, S. 1997. “Research, Patenting, and Technological Change.”Econometrica 65 (6): 1389–1419.
Kortum, S., and Lerner J. 2000. “Assessing the Contribution of Venture Capital to Innovation.”RAND Journal of Economics
31 (4): 674–692.
Lanjouw, J. O., and M. Schankerman. 2004. “Patent Quality and Research Productivity: Measuring Innovation with Multiple
Indicators.”The Economic Journal 114 (495): 441–465.
Lööf, H., and A. Heshmati. 2002. “Knowledge Capital and Performance Heterogeneity: A Firm-Level Innovation Study.”
International Journal of Production Economics 76 (1): 61–85.
Mansfield, E. 1980. “Basic Research and Productivity Increases in Manufacturing.”American Economic Review 70 (5): 863–873.
Pakes, A., and Z. Griliches. 1980. “Patents and R&D at the Firm Level: A first Look.”Economics Letters 5(4):377–381.
Also published in 1984 as 55–72 in Griliches, Z., R & D, Patents, and Productivity. University of Chicago Press.
Sakakibara, M., and L. Branstetter. 2001. “Do Stronger Patents Induce More Innovation? Evidence from the 1988 Japanese
Patent Law Reforms.”RAND Journal of Economics 32 (1): 77–100. Also published in 1999 as NBER Working Paper No. 7066.
Appendix 2: WoS subject field category distribution and mean impact of the papers
citing the 27 RIP core publications, across field classifications.
The layout of this map is based on a cluster analysis of citation links between subject field categories and grouped into 19
broader subject areas indicated on the map.
ECONOMICS OF INNOVATION AND NEW TECHNOLOGY 19
Downloaded by [Kungliga Tekniska Hogskola] at 09:59 06 November 2017
Legend: The size of circles reflects the number of papers classified as belonging to each category. The colour of the circles
reflects mean citation impact of all papers in that category (the scale on the left side). The numbers in/beside the larger
circles refer to the enumeration of fields in Table 3.
Appendix 3: List of papers used for comparison of citation functions.
Lee, D. S., E. Moretti, and M. J. Butler. 2004. “Do Voters Affect or Elect Policies? Evidence from the U. S. House.”The Quar-
terly Journal of Economics 119 (3): 807–859.
Kogut, B., and S. J. Chang. 1996. “Platform Investments and Volatile Exchange Rates: Direct Investment in the U.S. By Japa-
nese Electronic Companies.”The Review of Economics and Statistics 78 (2): 221–231.
Castelló, A., and R. Doménech. 2002. “Human Capital Inequality and Economic Growth: Some New Evidence.”The Econ-
omic Journal 112 (478): C187–C200.
Offerman, T., J. Sonnemans, and A. Schram. 1996. “Value Orientations, Expectations and Voluntary Contributions in Public
Goods.”The Economic Journal 106 (437): 817–845.
20 A. BROSTRÖM AND S. KARLSSON
Downloaded by [Kungliga Tekniska Hogskola] at 09:59 06 November 2017
