The Measurement of Innovation: A Systematic Review and Bibliometric Analysis of Global Innovation Index Research

Abstract

The purpose of this review is to synthesise the accumulated knowledge on Global Innovation Index (GII) research. We utilised a corpus from the Web of Science Core Collection to systematically examine the antecedents, consequences, and relationships among the GII’s dimensions. Additionally, we employed the bibliometric techniques of bibliographic coupling and co-citation analysis to identify the leading areas of GII research and the foundational literature in this field. Our systematic review of GII empirical research allowed us to graphically represent the significant relationships among its dimensions. The findings from the bibliographic coupling revealed five recent lines of investigation in GII research: configurational methods; innovation efficiency and policy; competitiveness, entrepreneurship, and sustainable development; innovation rankings; and culture. Furthermore, the co-citation analysis highlighted four clusters of literature that have contributed to GII research. We aim to enhance the field of Innovation Studies by showcasing the current state of research on the GII, one of the most promising tools for measuring innovation activity, and to provide insights into potential future research avenues to further develop this area of study.

1. Introduction

The influence of innovation on the competitiveness of firms and countries is increasingly acknowledged by scholars and policymakers (Fagerberg & Srholec, 2023; OECD, 2015). To inform policy development and to assess its impacts, innovation needs to be effectively measured (Gault, 2018). In this regard, the Oslo Manual (OECD/Eurostat, 2018, p. 247) defines innovation indicators as “a statistical summary measure of an innovation phenomenon (activity, output, expenditure, etc.) observed in a population or a sample thereof for a specific time or place. Indicators are usually corrected (or standardised) to permit comparisons across units that differ in size or other characteristics”. This general definition allows for various types of innovation indicators: direct measures of innovation (e.g., number of product/process innovations); direct measures of innovation activities (e.g., employees or expenditures in research and development (R&D)); and measures of framework conditions (e.g., infrastructure, tax incentives to invest in R&D).

Individual innovation indicators have been extensively used in empirical studies, such as sales of innovative products (Ebersberger et al., 2021), product or process innovations (Franco & Landini, 2022), R&D statistics (Coccia, 2018), or patent statistics (Burhan et al., 2017). However, they each present only a partial and often indirect view of the innovation phenomenon due to its intangible and not always directly observable nature (Grupp & Schubert, 2010). For instance, while patent data reflects an intention to protect the intellectual property of an idea, it does not inform about its actual implementation or commercialisation. Likewise, R&D expenditures reveal the efforts of a firm towards innovation but leave firms that do not conduct R&D unaccounted for.

In this context, indices and scoreboards have been developed by international organisations to capture the multidimensional nature of innovation and assess countries’ innovation panorama (Gault & Soete, 2022; Grupp & Schubert, 2010). Prominent examples are the Organisation for Economic Co-operation and Development (OECD) Innovation Indicators, the European Innovation Scoreboard (EIS), and the Global Innovation Index (GII).1

This paper focuses on the GII for several reasons. First, the GII provides a coherent organisation of its indicators into seven dimensions, explicitly distinguishing between innovation inputs and outputs. Second, it includes nearly 130 countries in each annual report, enabling a comprehensive worldwide comparison of innovation. Third, its framework is close to the key activities of innovation systems (Edquist, 2019). Fourth, it is becoming increasingly used in academic research, having been used to evaluate countries’ innovation performance (Fleacă et al., 2023) and efficiency (Erdin & Çağlar, 2023) and assess innovation determinants (Costa Cavalcante, 2024). A simple Google Scholar search of research published since 2020 reveals that the GII has almost four times the number of hits (16,800) than the EIS (4600); these are significantly larger numbers than those found in the Web of Science (WoS)—253 hits for the GII and 63 hits for the EIS—but of similar proportions.2

The GII adopts a clear input–output framework, akin to the rationale of the System of National Accounts (Godin, 2007), in the sense that innovation inputs allow for the existence of innovation outputs. The World Intellectual Property Organization (WIPO, 2023, p. 215) defines innovation inputs as the “elements of the economy that enable and facilitate innovation activities”, while defining innovation outputs as “the results of innovation activities within the economy”. However, how innovation inputs translate into innovation outputs is a matter of contested debate. While some authors see the input–output relationship as direct (Duarte & Carvalho, 2021; Oturakci, 2023), others view it as a complex network of structural relationships (Sohn et al., 2016), with some suggesting that innovation inputs combine to determine innovation outputs (Crespo & Crespo, 2016; Yu et al., 2021a). Furthermore, extant literature has proposed the existence of boundary conditions that could influence the innovation input–output relationship, such as national culture (Bendapudi et al., 2018) or income (Wang et al., 2021).

The extant literature led us to ask the following questions:

• RQ 1: What does the existing evidence tell us about the relationships between GII dimensions?
• RQ 2: Which antecedents and consequences of the GII have been analysed in the empirical literature?
• RQ 3: What can be expected of future GII research?

To answer these questions, an adequately conducted systematic literature review becomes necessary. Despite extensive research on this innovation indicator, no reviews have been conducted on GII research, limiting its role within the field of Innovation Studies.

Therefore, this systematic literature review aims to synthesise the accumulated knowledge on GII research. More specifically, we summarise the current empirical evidence on the relationships between GII dimensions and uncover the most recent lines of inquiry and the literature on which the GII research field is built, hence providing directions for future research. A search was conducted in the WoS Core Collection to obtain the corpus of GII research, which was analysed through the systematisation of information and bibliometric techniques (bibliometric coupling and co-citation). In doing so, this study contributes to the broader field of Innovation Studies. First, this is likely the first systematic review of GII research that consolidates existing knowledge and may guide future studies on national innovation. Second, this study clarifies the empirical relationships among the GII dimensions and their key antecedents and outcomes, enhancing researchers’ understanding of its mechanisms and reinforcing the index’s value for the field of Innovation Studies. Third, it informs future research through the identification of the most recent lines of investigation of GII research, highlighting possible paths for further inquiry.

The following section succinctly describes the GII, and then the methodology used is presented, including the collection of data and analyses made. Next, the results of the systematic review and bibliometric analyses are presented and synthesised, giving place to the formulation of future research directions, and, lastly, we conclude.

2. The Global Innovation Index (GII)

The GII is an annual report published since 2007 by the WIPO in collaboration with other organisations. The index measures national innovation, providing a detailed picture of the state of countries’ innovation performance and allowing the generation of meaningful comparisons (Dutta et al., 2007; WIPO, 2023).

The overall GII index is composed of two equally important sub-indices: the innovation inputs sub-index and the innovation outputs sub-index. The innovation inputs sub-index is composed of five pillars (institutions, human capital and research, infrastructure, market sophistication, and business sophistication), and the innovation outputs sub-index is composed of two pillars (knowledge and technology outputs and creative outputs). Each of the seven pillars is divided into three sub-pillars, each of which is composed of individual indicators. Figure 1 illustrates the pillars and sub-pillars of the GII, organising its 80 individual indicators along this framework.

The sub-pillars’ scores are calculated using a weighted average of normalised indicators, then normalised on a 0–100 scale to produce the pillars’ scores through another weighted average. The innovation input sub-index is the average of the five input pillars, while the innovation output sub-index is the average of the two output pillars. The overall GII score results from the simple average of both sub-indices. Data on individual indicators is gathered from various sources, such as the WIPO, the World Economic Forum’s Executive Opinion Survey, and the World Bank’s Worldwide Governance Indicators, among others. This results in a balanced mix of indicators, ranging from hard data or composite indicators to experts’ opinions.

3. Methods

This systematic review and bibliometric analysis were developed using the Preferred Reporting Items for Systematic Reviews and Meta-Analysis approach (Page et al., 2021).

3.1. Data Collection

Bibliographic data was collected through the WoS Core Collection database since it is considered a credible indexer of quality peer-reviewed journals (Huang et al., 2023; Maddi & Baudoin, 2022). More specifically, the Science Citation Index Expanded (1900–present), Social Sciences Citation Index (1956–present), and Emerging Sources Citation Index (2005–present) were chosen based on the subscription provided by our university. These sub-databases are reported to ensure the reproducibility of future studies (Liu, 2019).

We used the following query: TS = (“global innovation index”) OR TS = (“National innovation performance”) OR TS = (“national innovation capa*”). Additionally, we restricted the document type to “Article” and the language to “English”. The search period encompassed all articles published until the end of December 2023.

A total of 272 records were identified in the search. All identified records were carefully screened by reading their titles and abstracts, resulting in the exclusion of 71 records for addressing topics other than innovation. After retrieving the available full texts, we assessed the articles for their fit with the inclusion criteria:

• The article used the GII in at least part of its analyses;
• The article was not a literature review.

For instance, although they addressed national innovation, the article by Torres and Godinho (2023) was excluded because it employed alternative metrics of national innovation. Another example is the paper by Brás (2023), which is a data paper on the GII. After this step, we ended with a corpus of 156 articles, published in 97 journals between 2011 and 2023. Figure 2 shows the flow diagram of the article search.

3.2. Data Analysis

The raw data obtained from the search was imported into the bibliometrix R package (version 4.2.2) (Aria & Cuccurullo, 2017), allowing the automatic implementation of data cleaning procedures: capitalising all text; removing non-alphanumeric characters, punctuation symbols, and extra spaces; and truncating authors’ first and middle names to the first letter. Furthermore, we performed a manual disambiguation of all cited references and authors’ keywords. For example, Lundvall’s 1992 National Systems of Innovation could be labelled as Lundvall BA, 1992, National Systems of Innovation in one paper, and Lundvall B, 1992, National Systems of Innovation in another. Citations to annual reports, such as the GII or the Global Competitiveness Report (GCR), were considered as single documents (e.g., WIPO, GLOBAL INNOV INDEX for the GII, and WEF, GLOBAL COMPET REPORT for the GCR). The dataset used in this study is accessible at https://doi.org/10.5281/zenodo.15667102.

To answer our research questions, we analysed the data in two main steps. First, to answer RQ 1 and RQ 2, we developed a systematic review to identify and systematise the relationships between the GII dimensions based on the available evidence. Hence, we selected all papers from our corpus that studied the relationships between each dimension or the relationships between the GII or GII dimensions and other variables (46 papers). These papers were then manually analysed to identify significant relationships among the various innovation dimensions, as well as significant antecedents and outcomes.

Second, to answer RQ 3 and using the entire corpus, we conducted bibliographic coupling and co-citation analysis. Bibliometric methods are used to review a body of literature through statistical analysis of citation counts between publications (Culnan, 1986). Compared with other literature review techniques, bibliometrics provides analyses that are more reliable and more objective (Aria & Cuccurullo, 2017). Within the Innovation Studies literature, bibliometrics has been applied to ecosystem research (Shi et al., 2023), national systems of innovation (López-Rubio et al., 2022), technological catch-up (Souzanchi Kashani et al., 2022), and technology transfer (Bengoa et al., 2020). Each type of bibliometric method serves a purpose and has a distinct objective. Co-citation analysis clusters documents that are cited by two or more documents in the corpus (Small, 1973) and is usually applied to reveal the theoretical foundations of a field (López-Rubio et al., 2022); bibliographic coupling clusters documents in the corpus that share a large number of references (Kessler, 1963), making it suitable for uncovering research fronts (Boyack & Klavans, 2010); author co-citation is used to analyse the collaboration networks and knowledge exchanges within the field of research (White & Griffith, 1981); and co-word analysis clusters documents’ keywords to uncover the field’s conceptual structure (Callon et al., 1983).

Given the objectives of this study, bibliographic coupling and co-citation analysis were used to identify the frontier of GII research and the fields of research our corpus derives from, respectively. Hence, in the bibliographic coupling, the top 100 most connected documents were analysed using the Louvain community detection algorithm (Blondel et al., 2008), and their impact was measured by local citation score; clusters were labelled by approximation of the authors’ keywords. For the co-citation analysis, only documents (nodes) with a minimum of five connections (edges) were analysed using the Louvain algorithm. Both bibliographic coupling and co-citation analysis were mapped using biblioshiny (Aria & Cuccurullo, 2017). To obtain a detailed view of the recent lines of inquiry of GII research, we conducted a manual review of the 100 most connected papers identified by the bibliographic coupling. Each paper was then coded for its main theme and grouped by cluster. A similar procedure was developed for the co-citation analysis.

3.3. Overview of GII Research

Even though the GII was launched in 2007, the first report detailing countries’ scores on the various indicators was released in 2009. While some conference proceedings hinted at the usage of GII data in 2010, the first published article containing the GII in its analysis was developed by Iqbal (2011), where the authors used the GII for cross-country comparison.

Its use as an innovation indicator has been growing ever since (see Appendix A, Figure A1), a tendency partially explained by the overall growth of publications in the WoS (Liu et al., 2024). This growing number of papers has been published across various disciplines, including business, management, economics, and social sciences, as well as specialised journals on innovation. Sustainability appears as the journal with the highest number of GII papers in our corpus, followed by Marketing and Management of Innovations, Journal of Business Research, and Technological Forecasting and Social Change (see Appendix B,

Table A1. GII research publications from 2011 to 2023.

Sources	Articles
Sustainability	12
Marketing and Management of Innovations	7
Journal of Business Research	6
Technological Forecasting and Social Change	5
Economies	4
Entrepreneurship and Sustainability Issues	4
Financial and Credit Activity: Problems of Theory and Practice	4
Journal of Science and Technology Policy Management	4
Technology Analysis & Strategic Management	4
Cross-Cultural Research	3
Innovation & Management Review	3
Journal of the Knowledge Economy	3
Technology in Society	3

Note: Only journals with three or more articles are presented.

).

A concern exists regarding the quality of journals in our corpus, since close to 56% of the papers (n = 87) were published in journals indexed in the Emerging Sources Citation Index. This could indicate that the GII, although a useful and used tool for policy decision-making (Gault, 2023), lacks appeal to higher-quality scientific research. A grand limitation of the index for scientific research comes from the fact that a longitudinal approach is rather difficult to apply (Duarte & Carvalho, 2021) because the individual indicators and countries covered may change from one year to the next, rendering year-on-year comparisons unreliable. Nevertheless, Figure A1 reveals a trend of increasing usage of GII, since 2020, in journals indexed in the Social Sciences Citation Index and the Science Citation Index Expanded, which highlights the increased recognition of the GII in sound scientific research within the field of Innovation Studies.

4. Systematic Review

Our systematic review enabled us to address RQ 1 and RQ 2 by identifying the antecedents and consequences of the GII while also systematising the relationships among all dimensions of the GII. Antecedents were grouped into institutional, cultural, financial, innovative development, and other factors, while consequences were grouped into economic consequences, sustainability consequences, labour market consequences, and consequences for businesses. Figure 3 depicts the “antecedents–elements–consequences” framework of GII research (see Appendix C,

Table A2. Antecedents of the GII and its dimensions.

Antecedents	GII Variable	Suggested Relationship	Sources
Institutional factors
EFW	GII	+	Boudreaux (2017)
EFW	KTO	+	Boudreaux (2017)
EFW	CO	+	Boudreaux (2017)
Property rights	CO	+	Boudreaux (2017)
Regulation	CO	+	Boudreaux (2017)
Free trade	KTO	+	Boudreaux (2017)
CPI	GII	+	DiRienzo and Das (2015)
Public policies	GII	+	Kawabata and Camargo Junior (2020)
Real democracy	GII	+	Kawabata and Camargo Junior (2020)
Political trust	GII	+	Kawabata and Camargo Junior (2020)
Cultural factors
PDI	GII	−	Espig et al. (2022), Halushka et al. (2022), Kapoor et al. (2021), Lee et al. (2022), Rinne et al. (2012)
PDI	GII Outputs	−	Bukowski and Rudnicki (2019), Kapoor et al. (2021)
PDI	GII Inputs	−	Kapoor et al. (2021)
IDV	GII	+	Espig et al. (2022), Halushka et al. (2022), Kapoor et al. (2021), Lee et al. (2022), Lourenço and Santos (2023), Rinne et al. (2012)
IDV	GII Outputs	+	Bukowski and Rudnicki (2019), Kapoor et al. (2021)
IDV	GII Inputs	+	Kapoor et al. (2021)
MAS	GII	−−	Espig et al. (2022)
UAI	GII	+	Das (2022)
UAI	GII	−	Espig et al. (2022), Halushka et al. (2022), Lee et al. (2022), Lourenço and Santos (2023)
LTO	GII	+	Espig et al. (2022), Lee et al. (2022), Lourenço and Santos (2023)
LTO	GII Outputs	+	Bukowski and Rudnicki (2019)
IVR	GII	+	Espig et al. (2022), Lee et al. (2022), Lourenço and Santos (2023)
IVR	GII Outputs	+	Bukowski and Rudnicki (2019)
Autonomy	GII Outputs	+	B. Zhu et al. (2018)
Societal trust	GII Outputs	+	B. Zhu et al. (2018)
Self-protective values	KTO	−	Bendapudi et al. (2018)
Self-protective values	CO	−	Bendapudi et al. (2018)
Self-expansive values	KTO	+	Bendapudi et al. (2018)
Self-expansive values	CO	+	Bendapudi et al. (2018)
Ethnic diversity	GII	−	DiRienzo and Das (2015)
Religious diversity	GII	+	DiRienzo and Das (2015)
Financial factors
Domestic credit to the private sector	GII efficiency	−	Low et al. (2018)
Domestic stocks traded	GII efficiency	+	Low et al. (2018)
Market capitalisation	GII efficiency	+	Low et al. (2018)
Financial institution development	GII	+	Mursalov (2020)
Financial market development	GII	+	Mursalov (2020)
Revenue decentralisation	GII	+	Molotok (2020)
Innovation related factors
DESI	GII	+	Marino and Pariso (2021)
R&D expenditures	GII	+	Nabieva et al. (2021)
Cost of innovative goods/services	GII	−	Nabieva et al. (2021)
% innovative goods/services	GII	+	Nabieva et al. (2021)
Innovative entities	GII	+	Nabieva et al. (2021)
Other factors
Female labour participation	GII	+	Starchenko (2020)
Female labour participation with basic education	GII	+	Starchenko (2020)
Female unemployment	GII	+	Starchenko (2020)
Openness to experience	GII	+	Steel et al. (2012)
Agreeableness	GII	+	Steel et al. (2012)
GDPpc	GII Inputs	∩	Thangavelu et al. (2022)
GCI	GII	+	De Miranda et al. (2021)
Human Capital	GII	+	Suseno et al. (2020)
Social Capital	GII	+	Suseno et al. (2020)
PISA scores	KTO	+	Bendapudi et al. (2018)
PISA scores	CO	+	Bendapudi et al. (2018)
Configurational approaches
Outcome: GII outputs	Configurations: pdi*IDV + IDV*mas*UAI + IDV*UAI*LTO + PDI*idv*MAS*uai*LTO		Tekic and Tekic (2021)

Note: In configurations, lowercase represents the absence of a condition and uppercase represents its presence, * represents the Boolean operator “OR” and + represents the Boolean operator “AND”. +: positive relationship. −: negative relationship. ∩: inverse-U relationship. GII: Global Innovation Index. GII efficiency: ratio of GII outputs over GII inputs. KTO: Knowledge and Technology Outputs. CO: Creative Outputs. EFW: Economic Freedom of the World. CPI: Corruption Perception Index. PDI: Power Distance. IDV: Individualism. MAS: Masculinity. UAI: Uncertainty Avoidance. LTO: Long-Term Orientation. IVR: Indulgence. DESI: Digital Economy and Society Index. GDPpc: Gross Domestic Product per capita. GCI: Global Competitiveness Index. PISA: Programme for International Student Assessment.

,

Table A3. Consequences of the GII and its dimensions.

Consequences	GII Variable	Suggested Relationship	Sources
Economic consequences
GDPpc	GII	+	Dempere et al. (2023)
GDPpc	GII Inputs	+	Dempere et al. (2023)
GDPpc	GII Outputs	+	Dempere et al. (2023)
GDPpc	ITT	+	Dempere et al. (2023)
GDPpc	HCR	+	Dempere et al. (2023)
GDPpc	INF	+	Dempere et al. (2023)
GDPpc	BS	+	Dempere et al. (2023)
GDPpc	KTO	+	Dempere et al. (2023)
GDPpc	CO	+	Dempere et al. (2023)
GDP	GII	+	Çemberci et al. (2022)
FDI	GII	+	Çemberci et al. (2022)
Sustainability consequences
SDG Index	ITT	+	Chaparro-Banegas et al. (2023)
SDG Index	HCR	+	Chaparro-Banegas et al. (2023)
EPI	ITT	+	Fernandes et al. (2022)
EPI	HCR	+	Fernandes et al. (2022)
EPI	INF	+	Fernandes et al. (2022)
EPI	MS	−	Fernandes et al. (2022)
Labour market consequences
Self-employment	GII	−	Dempere et al. (2023)
Self-employment	GII Inputs	−	Dempere et al. (2023)
Self-employment	GII Outputs	−	Dempere et al. (2023)
Self-employment	ITT	−	Dempere et al. (2023)
Self-employment	HCR	−	Dempere et al. (2023)
Self-employment	INF	−	Dempere et al. (2023)
Self-employment	KTO	−	Dempere et al. (2023)
Self-employment	CO	−	Dempere et al. (2023)
Self-employment in the cultural sector	GII	+	Ljevak Lebeda et al. (2021)
Highly skilled migration	ITT	−	Labrianidis et al. (2023)
Highly skilled migration	BS	+	Labrianidis et al. (2023)
Highly skilled migration	KTO	+	Labrianidis et al. (2023)
Highly skilled migration	CO	−	Labrianidis et al. (2023)
Consequences for businesses
E-commerce capacity	HCR	+	Zygiaris (2022)
E-commerce capacity	INF	+	Zygiaris (2022)
E-commerce capacity	MS	+	Zygiaris (2022)
External knowledge acquisition	MS	−	Yurevich et al. (2023)
Business digital integration	ITT	+	Ionescu et al. (2022)
Firms’ average innovation performance	MS	+	Okatan and Alankuş (2019)
Firms’ appearance on innovation rankings	MS	+	Okatan and Alankuş (2019)
Configurational approaches
Outcome: GDP growth	Configurations: ITT*HCR*INF*MS + HCR*INF*MS*BS		Yu et al. (2021b)

Note: In configurations, uppercase represents its presence, * represents the Boolean operator “OR” and + represents the Boolean operator “AND”. +: positive relationship. −: negative relationship. GII: Global Innovation Index. KTO: Knowledge and Technology Outputs. CO: Creative Outputs. ITT: Institutions. HCR: Human Capital and Research. INF: Infrastructure. MS: Market Sophistication. BS: Business Sophistication. GDPpc: Gross Domestic Product per capita. GDP: Gross Domestic Product. FDI: Foreign Direct Investment. SDG: Sustainable Development Goals. EPI: Environmental Performance Index.

and

Table A4. Relationships between GII variables.

Independent	Dependent	Suggested Relationship	Sources
HCR	GII Outputs	+	Costa Cavalcante (2024)
INF	GII Outputs	+	Costa Cavalcante (2024)
MS	GII Outputs	−	Costa Cavalcante (2024)
BS	GII Outputs	+	Costa Cavalcante (2024), Duarte and Carvalho (2021)
ITT	GII	−	Oturakci (2023)
HCR	GII	+	Oturakci (2023)
INF	GII	−	Oturakci (2023)
MS	GII	+	Oturakci (2023)
BS	GII	+	Oturakci (2023)
KTO	GII	+	Oturakci (2023)
CO	GII	+	Oturakci (2023)
Configurational approaches
Outcome	Configurations		Source
GII	ITT*HCR*INF*BS*KTO*CO + ITT*hcr*INF*ms*bs*kto*co		Yu et al. (2021a)
GII Outputs	ITT*HCR*INF*BS + HCR*INF*MS*BS		Khedhaouria and Thurik (2017)
GII Outputs	High-income countries: INF + HCR + ITT*BS + ITT*MS Low-income countries: HCR*INF*BS + ITT*INF*BS + itt*HCR*MS*BS + HCR*INF*MS + ITT*INF*MS		Crespo and Crespo (2016)
GII Outputs	High-income countries: ITT*HCR*INF*BS + itt*hcr*INF*MS*bs + ITT*hcr*INF*ms*bs Upper-middle-income countries: ITT*INF*ms*BS + itt*HCR*MS*BS + ITT*HCR*INF*BS Lower-middle-income countries: ITT*hcr*inf*MS*bs + itt*HCR*inf*ms*BS		Wang et al. (2021)
KTO	ITT*HCR*INF*BS		Yu and Huarng (2023)
CO	ITT*HCR*INF*BS		Yu and Huarng (2023)

Note: In configurations, lowercase represents the absence of a condition and uppercase represents its presence, * represents the Boolean operator “OR” and + represents the Boolean operator “AND”. +: positive relationship. −: negative relationship. GII: Global Innovation Index. KTO: Knowledge and Technology Outputs. CO: Creative Outputs. ITT: Institutions. HCR: Human Capital and Research. INF: Infrastructure. MS: Market Sophistication. BS: Business Sophistication.

for detailed descriptions).

4.1. Antecedents

Cultural antecedents of innovation are the most studied factors in our corpus, with Hofstede’s cultural dimensions being the most utilised cultural framework. The reviewed literature agrees on the positive effects of low power distance, individualism, femininity, long-term orientation, and indulgence on innovation (Bukowski & Rudnicki, 2019; Espig et al., 2022; Kapoor et al., 2021; Rinne et al., 2012). Based on a configurational view of cultural dimensions, Tekic and Tekic (2021) identified equifinal cultural profiles that produce high GII outputs, highlighting the significance of collectivism and high power distance in these configurations.

Nevertheless, the role of uncertainty avoidance is still not consensual, with some studies finding a positive influence on the GII (Halushka et al., 2022; Lourenço & Santos, 2023) and others a negative influence (Das, 2022). This lack of consensus highlights different possible mechanisms through which uncertainty avoidance influences countries’ innovativeness. For example, it can be argued that people in countries with lower levels of uncertainty avoidance are more tolerant of risk-taking, which is a hallmark of entrepreneurs (Hébert & Link, 2006); hence, such countries would display higher levels of innovation. However, a counterpoint could be that societies more prone to high uncertainty avoidance levels are more likely to establish formal rules and structures to minimise the uncertainty levels (Hofstede et al., 2010), which may lead to a strengthened institutional setup within their national system of innovation (NSI) and hence better innovation performance.

The reviewed literature reports expected results regarding institutional and innovation-related factors. In particular, institutional quality tends to be associated with higher levels of innovation (Boudreaux, 2017; DiRienzo & Das, 2015; Kawabata & Camargo Junior, 2020), and comparable results were reported for the level of digital transformation and firms’ innovation activities (Marino & Pariso, 2021; Nabieva et al., 2021). Regarding financial factors, a higher level of financial development is positively associated with innovation (Low et al., 2018; Mursalov, 2020). However, Low et al. (2018, p. 516) suggest that the banking sector could be detrimental to countries’ efficiency in transforming innovation inputs into outputs, arguing that “powerful banks through their information production can somewhat hinder innovation by obtaining informational rents”.

Other factors influencing countries’ innovation levels have generally positive effects. (Bendapudi et al., 2018; De Miranda et al., 2021; Suseno et al., 2020). Of relevance, Thangavelu et al. (2022) found an inverted-U relationship between income and innovation inputs, indicating that low-income countries benefit more from increases in GDP per capita due to innovation than high-income countries. This finding is somewhat intuitive and likely linked to the principle of diminishing returns to scale.

4.2. Consequences

The reviewed literature supports economic growth theories’ claims about the role of innovation in the economic development of countries. For instance, Dempere et al. (2023) suggest that countries’ innovation levels are positively associated with income. Likewise, Çemberci et al. (2022) found that both GDP and FDI are influenced by countries’ innovativeness.

Concerning sustainability consequences, it has been found that institutions and human capital and research are important factors for the achievement of Sustainable Development Goals (Chaparro-Banegas et al., 2023). While these two factors, associated with good infrastructure, also benefit countries’ environmental performance, having a sophisticated market is seen as an environmental hindrance (Fernandes et al., 2022). This finding highlights the need to rethink the current economic paradigm by adopting sustainable consumption behaviours and business practices.

Regarding labour market consequences, Dempere et al. (2023) found negative relationships between innovation and self-employment, agreeing with entrepreneurship literature that argues that innovation fosters formal employment (Wennekers et al., 2005). However, evidence shows that this aggregated effect might be concealing specific sectoral nuances, for Ljevak Lebeda et al. (2021) revealed a positive connection between innovation and self-employment in the cultural sector. The authors recognise that this finding could be biased due to the COVID-19 pandemic, where cultural sector workers could have turned to self-employment in order to ensure their subsistence minimum.

The impact of innovation on the studied business outcomes was generally positive (Okatan & Alankuş, 2019; Zygiaris, 2022), except for market sophistication on firms’ external knowledge acquisition (Yurevich et al., 2023).

4.3. Relationships Between GII Dimensions

Conceptually, one would expect innovation input to be positively associated with innovation outputs or the overall GII. However, the evidence suggests negative relations between market sophistication and innovation (Costa Cavalcante, 2024) and between institutions and infrastructure and the overall GII (Oturakci, 2023). Configurational approaches, by considering equifinality, help understand these findings by providing evidence of different combinations of inputs that lead to high GII outputs (Khedhaouria & Thurik, 2017; Yu & Huarng, 2023). For instance, Khedhaouria and Thurik (2017) revealed that market sophistication was indifferent to high GII outputs in one combination where all other inputs were present, and that the institutional dimension was indifferent in the presence of all other inputs in another combination. Furthermore, both Crespo and Crespo (2016) and Wang et al. (2021) demonstrated that the level of national income matters for the arrangements of inputs in achieving high innovation outputs, suggesting a possible moderating effect.

Figure 4 represents a systematisation of the relationships between GII dimensions, as analysed in the corpus, in which the articles are based on regression-based methods and structural equation modelling (Costa Cavalcante, 2024; Sohn et al., 2016). In the figure, arrows represent the significant relationships found among the dimensions, and signs represent the direction of the effect. Relationships are mostly positive, except for institutions and knowledge and technology outputs, whereas conflicting findings were observed regarding the effects of market sophistication (Costa Cavalcante, 2024; Sohn et al., 2016).

The inconsistent results found regarding the market sophistication dimension are likely derived from the methodological limitations of the GII, namely the absence of a harmonised longitudinal database that would allow for comparability across time. Evidence supporting this argument comes from Duarte and Carvalho (2021), who developed a longitudinal version of the GII and incorporated a temporal lag in inputs, having observed a change in sign for the effect of market sophistication on innovation outputs. The authors found that adding a 1-year time lag to market sophistication changed the estimated sign from negative to positive, although this remained a non-significant relationship. This evidence indicates that research using the GII should find ways to incorporate a temporal dimension, as innovation inputs are unlikely to translate into innovation outputs instantaneously.

5. Bibliometric Coupling

Bibliometric coupling analysis uncovered five clusters (Figure 5) representing the most recent lines of investigation of GII research, which allowed us to answer RQ 3.

Table 1. Clusters and main themes of the most recent lines of GII research.

Cluster	Main Themes	Key Papers
Configurational methods (red)	Qualitative comparative analysis	Crespo and Crespo (2016), Ding (2022), Huarng and Yu (2022), Khedhaouria and Thurik (2017), Yu and Huarng (2023), Yu et al. (2021a)
Innovation efficiency and innovation policy (blue)	Innovation efficiency	Bakhtiar et al. (2022), Erdin and Çağlar (2023), Jankowska et al. (2017), Tziogkidis et al. (2020)
	Innovation policy	Duarte and Carvalho (2021), Fakhimi and Miremadi (2022), P. Jackson et al. (2016), Kuzior et al. (2022)
Competitiveness, entrepreneurship, and sustainable development (green)	Competitiveness	Cahyadi and Magda (2021), De Miranda et al. (2021), Tarnenko (2013), Varaksina et al. (2022)
	Entrepreneurship	Albulescu and Drăghici (2016), Dempere et al. (2023), Heidor et al. (2022), Nazarov et al. (2022)
	Sustainable development	Chaparro-Banegas et al. (2023), Fonseca and Lima (2015), Ibraghimov (2022)
Innovation rankings (purple)	Innovation rankings	Alidrisi (2021), Aytekin et al. (2022), Pençe et al. (2019), Saisse and Lima (2019)
Culture (yellow)	Cultural values	Bendapudi et al. (2018), Bukowski and Rudnicki (2019), Espig et al. (2022), Lee et al. (2022), Kapoor et al. (2021), Rinne et al. (2012), B. Zhu et al. (2018)

summarises the cluster labelling, the main themes of each cluster, and their most representative papers.

The “Competitiveness, entrepreneurship, and sustainable development” cluster (green) is the largest with 55 articles, followed by the “Innovation efficiency and innovation policy” cluster (blue—15 articles) and the “Innovation rankings” cluster (purple—14 articles). The “Culture” (yellow) and “Configurational methods” (red) clusters encompass 8 articles each. The average year of publication for the articles in each cluster is as follows:

• Red: 2020.6
• Blue: 2019.6
• Green: 2019.9
• Purple: 2021.1
• Yellow: 2018.8

This reveals the recency of articles addressing countries’ innovation rankings and the uptake of configurational research methods in GII studies. In the remainder of this section, we discuss the main themes of research within each cluster.

5.1. Cluster 1: Configurational Methods

Table 1 shows that the first cluster relates to configurational methods. Much of this line of research has focused on combinations of the GII dimensions, exploring questions such as how innovation inputs combine to explain innovation outputs, and how combinations of innovation inputs vary in explaining the innovation outputs of countries at different levels of development.

A common finding in these studies relates to the dimension of market sophistication, where its presence or absence, in combination with the presence of the other innovation inputs, turns out to be indifferent for high innovation levels (Huarng & Yu, 2022; Khedhaouria & Thurik, 2017; Yu & Huarng, 2023). Nevertheless, Crespo and Crespo (2016) showed that combinations of innovation inputs leading to high innovation outputs differ across levels of economic development, suggesting that the innovation input–output relationship is likely conditioned by national income levels.

5.2. Cluster 2: Innovation Efficiency and Innovation Policy

Innovation efficiency and innovation policy are the main themes identified in the second cluster. GII research on innovation policy typically draws from NSI approaches to make cross-country comparisons to gather insights on key policy areas that should be addressed to spur innovation (Duarte & Carvalho, 2021; Kuzior et al., 2022). Relevant policy areas highlighted by works in this cluster include the development of human capital (P. Jackson et al., 2016; Kuzior et al., 2022), encouraging inward FDI (Duarte & Carvalho, 2021), or bolstering the interactions between academic and industrial sectors (Fakhimi & Miremadi, 2022).

Regarding innovation efficiency analyses, the main assumption is that some countries may be more efficient in transforming innovation inputs into innovation outputs than others (Erdin & Çağlar, 2023). In this context, data envelopment analysis (DEA) has been used to effectively measure countries’ innovation efficiency, revealing how investments in innovation translate into actual outputs (Erdin & Çağlar, 2023; Tziogkidis et al., 2020). For instance, Tziogkidis et al. (2020) found substantial diversity in countries’ NSI. In particular, the authors observed that low-income countries were significantly ineffective in obtaining creative outputs compared with knowledge and technology outputs, which could reflect national priorities towards innovation. Furthermore, European high-income countries showed better efficiency rates than high-income countries elsewhere, stressing the role of regional characteristics and market openness to countries’ innovation.

5.3. Cluster 3: Competitiveness, Entrepreneurship, and Sustainability

Cluster 3 encompasses works devoted to the main themes of competitiveness, entrepreneurship, and sustainability. Papers within the competitiveness theme try to uncover the innovation determinants of competitiveness (Cahyadi & Magda, 2021), as well as the competitiveness factors that lead to innovation (Varaksina et al., 2022). These studies typically focus on specific groups of countries, such as the G20 (Cahyadi & Magda, 2021) or Central and Eastern European countries (Varaksina et al., 2022), with a notable exception that includes a large number of economies (De Miranda et al., 2021). The main finding of this line of research is a clear connection between national innovation and competitiveness (De Miranda et al., 2021; Varaksina et al., 2022). More specifically, the dimensions of general infrastructure, intangible assets, ecological sustainability, investment, and information and communication technologies were found to be the most important drivers of countries’ competitiveness, while higher education and training were the most salient competitiveness factors influencing national innovation (De Miranda et al., 2021; Tarnenko, 2013).

Regarding entrepreneurship, studies draw from the notion that entrepreneurial activity is a key factor in countries’ innovation levels (Dempere et al., 2023; Nazarov et al., 2022). Findings, nevertheless, highlight a more nuanced explanation of how entrepreneurship influences national innovation. For instance, Nazarov et al. (2022) found that venture capital availability positively affected national innovation, while the number of new firms and amount of minority investor protection revealed a negative effect. Albulescu and Drăghici (2016), on the other hand, found no relationship between total entrepreneurial activity, or the number of innovation-driven entrepreneurs, and national innovation. These findings hint at the possible influence of contextual factors in shaping this relationship, whether institutional or cultural factors or specific national policies to stimulate entrepreneurship.

The topic of sustainable development looks mostly at the connections between national innovation and the achievement of the Sustainable Development Goals (SDG) (Chaparro-Banegas et al., 2023; Ibraghimov, 2022) or social sustainability (Fonseca & Lima, 2015). Findings seem to indicate a positive impact of innovation on sustainable development (Chaparro-Banegas et al., 2023; Fonseca & Lima, 2015), with a single study indicating a negative correlation between national innovation and the achievement of SDGs (Ibraghimov, 2022), though it lacked a robustness analysis to verify such a finding.

5.4. Cluster 4: Innovation Rankings

The fourth cluster identified relates to the development of new country rankings based on their innovation levels and other criteria (Aytekin et al., 2022; Pençe et al., 2019). Various methods are used for this purpose (e.g., artificial neural networks, TOPSIS), though DEA is the most common approach (Alidrisi, 2021; Aytekin et al., 2022). Since data is usually modelled from several inputs and outputs, the GII becomes a suitable candidate for this kind of analysis.

The main findings in rankings were usually similar but dependent on the sample of countries being analysed. Analyses of European countries normally showed Nordic countries, Germany, and the Netherlands in the top spots (Aytekin et al., 2022; Marti & Puertas, 2023), while worldwide analyses revealed the dominance of Japan, the US, the UK, Germany, and the Republic of Korea (Alidrisi, 2021; Barragán-Ocaña et al., 2020).

Insights from this particular field could contribute to improvements in the GII itself, possibly enhancing the way countries are ranked in the index.

5.5. Cluster 5: Culture

The fifth cluster identified is dedicated to the investigation of the relationship between national culture and innovation. Research in this field mainly argues that societies with certain cultural values have a higher propensity to innovate than others (Bendapudi et al., 2018; Rinne et al., 2012). Besides the dominance of Hofstede’s cultural measurements, a few studies (Bendapudi et al., 2018; B. Zhu et al., 2018) use GLOBE’s (House et al., 2004) or Schwartz’s (1994) cultural frameworks.

Consistent with the seminal work of Shane (1992), culture–innovation research has shown the importance of low power distance and individualism to national innovation (Bukowski & Rudnicki, 2019; Lee et al., 2022; Rinne et al., 2012). Nevertheless, evidence of other cultural dimensions’ effects on innovation has also been reported, such as femininity, uncertainty acceptance, long-term orientation, and indulgence (Bukowski & Rudnicki, 2019; Espig et al., 2022; Lee et al., 2022). Nevertheless, recent research has proposed that technology might be influencing cultural change (Salehan et al., 2018).

Some researchers question whether contextual factors moderate or mediate the relationship between culture and innovation (Kapoor et al., 2021; Lee et al., 2022), or even if national culture takes on a moderator role in national innovation processes (Bendapudi et al., 2018). For instance, Kapoor et al. (2021) found that in individualistic and low-power-distance countries, lower innovation was associated with more governmental stringency during COVID-19. However, Lee et al. (2022) found no differences in the cultural profile required for high innovation inputs and outputs before or after the crisis period, yet they found national income as a moderating variable. Bendapudi et al. (2018), on the other hand, discovered that a high level of self-protective values (power, achievement, conformity, tradition, and security) hinders the relationship between basic education and creative outputs. This literature suggests that research questions involving culture as a boundary condition for the transformation of innovation inputs into outputs might provide fruitful contributions.

6. Co-Citation Analysis

The co-citation analysis revealed four clusters (Figure 6) representing the knowledge bases used by the studies of our corpus. The identified communities were labelled “Economics literature” (red), the largest cluster with 19 articles; “Innovation systems” (green—10 articles); “Configurational methods” (blue—9 articles); and “Culture” (purple—6 articles). In the remainder of this section, we illustrate each cluster and its representative works (

Table 2. Clusters and main themes of knowledge bases of GII research.

Cluster	Main Themes	Key Papers
Economics literature (red)	Economic growth theory	Lucas (1988), Romer (1986, 1990), Schumpeter (1934), Solow (1956)
	National innovation capabilities	Fagerberg and Srholec (2008), Furman et al. (2002), Lundvall et al. (2002)
Configurational methods (blue)	QCA theory	Pappas and Woodside (2021), Ragin (2008), Woodside (2013)
	GII applications	Crespo and Crespo (2016), Khedhaouria and Thurik (2017)
Innovation systems (green)	National systems of innovation	Edquist (2006), Freeman (1987, 1995), Lundvall (1992, 2007), Nelson (1993)
	Innovation systems in developing countries	Metcalfe and Ramlogan (2008), Watkins et al. (2015)
Culture (purple)	Cultural values	Hofstede (2001), Hofstede et al. (2010), Rinne et al. (2012), Shane (1992, 1993), Taylor and Wilson (2012)

).

6.1. Cluster 1: Economics Literature

The first cluster concerns economic literature, revealing that economic growth theories are the ground on which GII research developed as a particular field of Innovation Studies. Within this field of research, the GII report is, understandably, the most co-cited work, being in close connection with the Global Competitiveness Report and the Oslo Manual.

The main themes identified in this cluster are economic growth theory, based on the papers by Lucas (1988), Romer (1986, 1990), and Solow (1956), and national innovation capabilities, based on the papers by Fagerberg and Srholec (2008), Lundvall (2010), and Lundvall et al. (2002). This cluster sets out the foundation of most GII research by highlighting the role of technological change and innovation in the economic development of countries (Romer, 1990; Schumpeter, 1934; Solow, 1956). Economic growth theories laid the foundations for investigating the role of innovation in countries’ economic development (Çemberci et al., 2022; Vlasova & Saprykina, 2024). Schumpeter’s (1934) work is considered a significant contribution to Innovation Studies, highlighting the importance of entrepreneurship and innovation in driving economic growth. Even though considerations about the influence of innovation on countries’ development were well established in the middle of the 20th century, official statistics on science and technology were only released later—primarily through the NBER, the National Science Foundation, and the OECD (Godin, 2007)—setting in motion the field of Innovation Studies.

Innovation capabilities within the NSI approach are also identified in this cluster. A common assumption of this line of inquiry is related to the capabilities needed to achieve higher economic development (Fagerberg & Srholec, 2008). Among such capabilities, market demand and its interactions with the production system play a prominent role in the development of innovations and economic growth (Furman et al., 2002; Lundvall et al., 2002). The main argument is that innovation occurs when the user sector interacts with the producer sector through relationships of coordination and cooperation, termed interactive learning (Lundvall et al., 2002). Furthermore, other capabilities, in the form of institutions, play a significant role in shaping these interactions (Fagerberg & Srholec, 2008; Lundvall et al., 2002). Both formal and informal institutions help determine the way people relate to each other and how knowledge is used in society (Lundvall et al., 2002).

6.2. Cluster 2: Configurational Methods

The second cluster identified is related to the knowledge bases for configurational approaches to innovation. Table 2 demonstrates that the main themes identified are the theoretical roots of qualitative comparative analysis and its applications within GII research.

Qualitative comparative analysis (QCA) is a methodological alternative in the social sciences to purely quantitative and qualitative approaches (Ragin, 2008). This method is based on set theory and investigates social phenomena through the discovery of set relationships. More specifically, QCA can identify statements that describe different combinations of conditions that are sufficient for a specific outcome to occur (Pappas & Woodside, 2021). Within GII research, adopting a configurational approach has become an important method of analysis, for it allows combinations of innovation inputs in determining countries’ innovation outputs (Crespo & Crespo, 2016; Khedhaouria & Thurik, 2017). This method fits well with the NSI approach due to the theoretical underpinning of interactions between the elements that constitute countries’ NSI, particularly the interactions among national institutions and innovation actors (Edquist, 2019).

6.3. Cluster 3: Innovation Systems

The third cluster identified is related to innovation systems literature, where we identified two main themes: national systems of innovation and innovation systems in developing countries. This cluster shares the lifelong work of Bengt-Åke Lundvall with the “economics literature” cluster, demonstrating the multidisciplinary nature of the field by crossing both economic development and innovation systems literatures.

This cluster is built on the early conceptualisations of innovation systems, put forth by the works of Freeman (1987), Lundvall (1992), and Nelson (1993). The NSI approach appeared as an opposing view of the new growth theory (Romer, 1990), by focusing on theories of interactive learning (Lundvall, 1992) and evolutionary economics (Nelson & Winter, 1982). The rationale for adopting the national level of analysis is grounded in the idea that culture, language, norms, policies, education, and many institutions have a national character (Lundvall, 1992). Thus, a major theme in NSI theorising revolves around the national actors involved in innovation development, their interactions, and the institutions that enable the emergence of innovations (Edquist, 2006).

While initial NSI studies provided a detailed description of innovation systems in developed countries (Freeman, 1987; Lundvall, 1992; Nelson, 1993), proposals for its adaptation to developing countries soon became apparent (Lundvall, 2007). Since developing countries lack a broad knowledge base from which local firms can draw, there are few alternatives other than relying on foreign knowledge and technology (Metcalfe & Ramlogan, 2008). Metcalfe and Ramlogan (2008) argue that a key challenge in applying an NSI framework to developing countries is that it prioritises activities at the technological frontier, neglecting the significance of continuous improvement and changes in organisational routines or production processes among domestic firms. Furthermore, Watkins et al. (2015) argued that the innovation drivers in developing countries might differ from those in developed nations, making a poorly adapted NSI analysis potentially unhelpful and even misleading for policymakers in developing countries.

6.4. Cluster 4: Culture

The fourth cluster identified relates to the knowledge bases of culture’s influence on innovation. Shane’s seminal works draw on Hofstede’s cultural values to explore why some countries innovate more than others. Cultural dimensions thought to be associated with innovation activities include power distance, individualism, and uncertainty avoidance (Shane, 1992, 1993). Given that high-power-distance societies exhibit a marked hierarchical structure with a bureaucratic focus (Hofstede, 1980), knowledge exchange and the necessary creativity for innovation could become hindered; thus, less power-distant societies would be more innovative (Shane, 1992, 1993).

Individualistic societies are outward-oriented and value freedom more than collectivistic societies (Hofstede, 1980). This is important for innovation because freedom to take action is necessary for entrepreneurship to happen, and an outward orientation facilitates contact with others, thus stimulating creativity (Shane, 1993). Even though Shane’s works confirmed the importance of individualistic societies to the existence of innovation, typical collectivistic countries (e.g., Southeast Asian countries) have been showing impressive rates of technological catch-up (Taylor & Wilson, 2012). Unsurprisingly, Taylor and Wilson (2012) and other recent research (Tekic & Tekic, 2021) have found evidence of an important contribution of collectivism to the development of innovation.

Given that innovation activities usually have uncertain outcomes, a degree of tolerance for uncertainty is likely to play a role in pursuing innovation (Shane, 1993). While Shane (1993) confirmed this to be an important determinant of innovation rates, Rinne et al. (2012) failed to find a significant contribution of uncertainty acceptance to countries’ innovation levels.

Overall, the corpus focuses heavily on innovation systems, economic development, and the role of institutions in shaping these processes. The network (Figure 6) highlights the influence of classical works in Innovation Studies (Freeman, 1987; Lundvall, 1992; Nelson, 1993; Schumpeter, 1934; Solow, 1956) alongside the contemporary relevance of organisations like the WIPO and the World Economic Forum (WEF). Ultimately, the presence of different communities suggests that the field is multidisciplinary, drawing from economics, management, sociology, and political science.

7. Directions for Future Research

GII research is a promising area within the field of Innovation Studies, but further exploration in various directions is needed. Findings from the systematic review, bibliometric coupling, and co-citation analysis led us to propose the following research directions.

7.1. Factors Affecting the Innovation Input–Output Relationship

The GII emulates innovation activities within the economy, and relationships between them are likely to differ, in degree and kind, from one country to another. Several studies have offered insights into the relationships between GII dimensions (Duarte & Carvalho, 2021; Fakhimi & Miremadi, 2022; Sohn et al., 2016), as well as how the economic development of countries shapes these relationships (Crespo & Crespo, 2016; Vlasova & Saprykina, 2024). However, economic development is but one of many factors that shape innovation processes. Hence, we offer two promising avenues to explore moderating factors in the innovation input–output relationship.

First, we demonstrated that cross-cultural research is a theoretical foundation that GII research often uses. At the same time, studies about the relationship between culture and innovation were found to be one of the most recent lines of inquiry of GII research, which is associated with many cultural factors considered as antecedents to countries’ innovation levels. Some of those lines of inquiry propose a moderating or mediating role of national culture (Kapoor et al., 2021; Lee et al., 2022; B. Zhu et al., 2018). Thus, we suggest that national culture, rather than a determining factor, could be a boundary condition that shapes how societies transform innovation inputs into outputs. Achieving innovation outputs requires interactions among different national actors (e.g., firms, academia, government) (Edquist, 2006; Lundvall, 2007), which, in turn, are developed within specific cultural environments. In this sense, there may be interactions occurring between cultural dimensions and innovation inputs. For instance, Bendapudi et al. (2018) showed that self-protective values dampened the relationship between quality basic education and creative outputs, but not knowledge and technology outputs. Further research in this area may delve deeper into the role played by culture in national innovation processes. It may also prove a fruitful avenue to examine the role of formal institutional arrangements in these innovation processes.

Second, while the use of configurational methods has been rising in GII research, more studies using this methodology are required. This is particularly important in cultural research since Hofstede (2011) stressed the importance of considering the six cultural dimensions in combination. Furthermore, both the literature on NSI (Lundvall, 2007) and that on the varieties of capitalism (G. Jackson & Deeg, 2019) posit that institutions, formal and informal, interact with one another in determining innovation. Recent methodological proposals have paved the way for analysing moderation (Ma et al., 2024) and mediation effects (Baumgartner, 2009) within the configurational environment, which could provide the necessary tools to further develop the GII research field.

7.2. GII Research in Developing Countries

Seminal NSI studies have mainly focused on developed countries from Europe, North America, and Japan (Freeman, 1987; Lundvall, 1992; Nelson, 1993), which created a need to adapt the systemic analyses to countries in the Global South (Arocena & Sutz, 2000; Lundvall, 2007). One of the themes emerging from co-citation analysis was that of innovation systems in developing countries, indicating that GII research draws from that literature to explore NSI in less developed countries. As a consequence, a recent line of investigation is dedicated to studying the impact of innovation on competitiveness, particularly in countries with lower income levels (Tarnenko, 2013; Varaksina et al., 2022).

Therefore, we find that the GII could be an adequate tool for analysing NSI in the Global South, for it does not rely exclusively on indicators of the technological frontier (e.g., patents, scientific publications) to depict the innovation activities within countries. For example, the GII has been used to assess low- and middle-income countries’ NSI, such as Syria (Alnafrah & Mouselli, 2020), Kazakhstan (Kurmanov et al., 2019), BRICS countries (Petkovski, 2023), and Ukraine (Kuzior et al., 2022). In this sense, further GII research is needed to uncover the uniqueness of developing countries’ NSI, so that effective innovation policies could be set in motion.

7.3. Methodological Improvements

The GII is a comprehensive innovation analysis tool, and each report is statistically validated by the European Commission’s Joint Research Centre (WIPO, 2023). Even though indicators are validated for their contribution to the index, the methodology adopted to construct the GII presents some constraints that may hinder its practical utility. A recent line of research has emerged that focuses on various methodologies to obtain more-accurate rankings of country-level innovation, namely artificial neural networks (Pençe et al., 2019) and DEA (Alidrisi, 2021). These and other methodological improvements lead us to suggest two potential future research directions.

First, countries’ innovation rankings influence governments and policymakers by adjusting their behaviour regarding adopting public policies to foster innovation (Gault & Soete, 2022). A particularly important question is which factors should be invested in more to climb the ranking leader (Aytekin et al., 2022). Although the GII ranks countries according to their overall innovation levels, it attributes the same importance to all innovation pillars, which obscures the most relevant factors for high innovation. Therefore, exploring efficiency-based ranking techniques is likely to uncover more-accurate rankings of countries’ innovativeness.

Second, the methodology used in constructing the GII renders year-on-year comparisons impossible (WIPO, 2023). Since the scores of each indicator are standardised according to the minimum and maximum values in each year, both scores and respective rankings become incomparable from one year to the next due to the inclusion/exclusion of countries or the inclusion/exclusion of indicators. Given the importance of longitudinal analysis to understand the dynamics of innovation, future studies should invest in the harmonisation of the index before constructing panels for longitudinal analysis (see Duarte and Carvalho (2021)).

8. Conclusions

The purpose of this study was to synthesise the accumulated knowledge on GII research. Following a systematic review of studies using the GII, we were able to systematise the evidence on the relationships between all dimensions of the GII, as well as the antecedents and consequences of the index. Furthermore, we identified five clusters of themes most recently investigated in GII research and four main clusters of the literature upon which this field builds, allowing us to propose directions for future studies. The novelty of this study rests on being the first systematic review of GII research. Thus, it contributes to consolidating the GII innovation indicator within the fields of Innovation Studies. Its unique input–output structure makes the index a tool capable of dealing with the interactions occurring between the various innovation activities. Moreover, its analytical framework allows integration in a variety of research questions, hence helping to expand innovation research.

As for the limitations of this study, the most significant is the reliance on a single database, the Web of Science. Previous research has found significant disparities in the coverage provided by WoS across regions, disciplines, and languages (Asubiaro et al., 2024; Singh et al., 2021). While most good-quality journals are indexed in the WoS, we recognise that good research may be published in journals outside the scope of the database. Furthermore, the WoS includes caveats regarding old literature retrieval and historical bibliometric analysis (Liu, 2021), such as incomplete records before 1990 and a lack of full authors’ names prior to 2006. These caveats could hinder the reproducibility of topic/keyword analysis using the WoS. Thus, future literature review studies may consider the use of different databases to complement the WoS. In this regard, Scopus could also be used since it is a widely used database (J. Zhu & Liu, 2020), or Ulrich’s Periodicals Directory, which is the most comprehensive database of journals published throughout the world (Asubiaro et al., 2024).