Does tertiary education promote technological innovation sustainability? The role of national intellectual capital. An empirical evidence

Abstract

Tertiary education redefines its role in research and innovation, acting as a proponent of the research and innovation culture, creating real-world solutions, and bridging the gap between decision-making, governance, and innovation. We empirically explore tertiary education’s effect on innovation and national intellectual capital’s role. The empirical findings of the 2SLS and instrumental variable fixed effect model with Driscoll-Kraay robust standard errors, based on panel evidence from 79 economies from 1995 to 2017, show that tertiary education has a positive and substantial effect on innovation performance as determined by patent and trademark applications. Comparable results are obtained using instrumental variable panel quantile regression as robustness tests to support the findings. Also, theoretically, the result shows that national intellectual capital reinforces the impact of tertiary education as a catalyst for technological progress. The findings support R&D data policies that support specific digital innovation, skills, knowledge creation, and diffusion. Proactive policy frameworks should be adopted to promote national intellectual capital through internet infrastructure for economic inclusion that fosters innovativeness; tertiary education using digital technology redefines creativity, stimulates cooperation, and aids in forming innovative ecosystems; boosting government and private grants to university academics allows for identifying ideas with the greatest long-term potential.

HIGHLIGHTS

• Tertiary education and technological progress in a panel of 79 economies from 1995–2017 were examined.

• Empirical results using 2SLS, IV fixed effect model with Driscoll-Kraay robust standard errors, and IV panel quantile regressions indicate tertiary education is essential to innovation.

• National intellectual capital reinforces the impact of tertiary education to promote technological progress.

• The findings support R&D data policies that support specific digital innovation, skills, knowledge creation, and diffusion.

• Proactive policy frameworks should be adopted to promote national intellectual capital through internet infrastructure for economic inclusion that fosters innovativeness.

• Tertiary education using digital technology can redefine creativity, stimulates cooperation, and aids in forming innovative ecosystems.

• Boosting government and private grants to university academics allows for identifying ideas with the greatest long-term potential.

Keywords:

• Tertiary education
• innovative activity
• research and development
• national intellectual capital

JEL CLASSIFICATION:

• I23; O31; O3

1. Introduction

The paper highlights the importance of tertiary education efficiency in the knowledge society as drivers of R&D and technological innovation (Tech.Inno). One of the main policy directions is to strengthen the role of tertiary learning in R&D and Tech.Inno performance – to improve knowledge dissemination, improve interaction channels, stimulate inter-institutional cooperation, and use tertiary education (Ter.Ed) industry to promote R&D internationalisation. The growth of the knowledge society is a direct result of a complex interplay of economic, social, and cultural factors that include knowledge generation and unbiased circulation, open access, and distribution (Muresan & Gogu, 2012). The overall approach of the knowledge-based economy aims at, first and foremost, the expansion and rising proliferation of knowledge-intensive products and services, which is the determining factor. As a result, the primary issue of the knowledge economy is the development of new competencies among global citizens, such as the ability to learn constantly and produce knowledge. In this framework, general and postsecondary education and lifelong learning are the primary pillars of the knowledge culture, alongside technical progress, research, and innovation. Furthermore, Ter.Ed is actively engaged in R&D and innovation, positioning itself as a key player pursuing sustainable development. In the current environment, universities are active advocates of R&D and Tech.Inno culture and active engines to create innovative solutions in technical and socio-economic fields.

Tech.Inno has become one of the greatest economic growth and development pathways. It stimulates other strategies and actions stirred toward economic growth and development. The functionality of other capital inputs depends on national intellectual capital (NIC) knowledge to function, which quickens other capitals into action and enhances them to become productive. Therefore, if a country’s innovative capacity must grow to increase economic growth and development, its NIC must be robust enough to create new ideas that cause change. Innovative capacity places the country at the forefront of market competition and equips the country with the required knowledge to develop new ideas and solutions to human problems (Pradhan et al., 2016). Therefore, Tech.Inno is a significant phenomenon and the major engine of economic progress, and the level and structure of innovation should always be considered for economic progress. In addition, NIC is regarded as a major driver for Tech.Inno, due to the knowledge, skills, and experiences it equips one with because NIC is identified as the amount of individual or collective knowledge used to generate additional forms of capital or upsurge their worth.

However, we focus on identifying the generic drive for innovation – the Ter.Ed becomes a thing of proper consideration because of its significant role in the production of ‘knowledge, skills, and experiences,’ which forms the basis for both NIC and R&D. As a result, we consider universities to be a pipeline of NIC since it is there that we create human capital and get the essential information about intellectual capital (IC) (Todericiu & Şerban, 2015). Therefore, this work seeks to study Ter.Ed’s influence on Tech.Inno and NIC's role on Tech.Inno performance. The research’s major outcome is a good repositioning of universities in the socio-economic environment. Accelerating and increasing the transfer of information and technology from the academy to the market might better exploit the research result, increase the cross-fertilisation of ideas, and quick implementation in the socio-economic environment. Strengthening the link between universities and the sphere of governance is another significant contribution that facilitates informed decision-making. Figure 1 below shows the gross enrollment ratio in Ter.Ed in different regions as of 2014. Figure 2 shows the worldwide Ter.Ed enrollment from 1970 to 2020. These figures indicate that there has been a somewhat increase in tertiary school enrolment across the globe.

Figure 1. Gross enrollment ratio in tertiary education, 2014.

Source: Unesco publication (2020).

Figure 2. Worldwide school enrollment, tertiary (% gross) from 1970 to 2020.

1.1. The role of tertiary education in innovation generation

Engaging acquired knowledge, skills, and experiences strategically produces technological innovation (Tech.Inno). However, a good environment full of ideas, skills, experiences, and knowledge must be implemented for this to happen. Hence, the development and reproduction of the innovative environment are determined by education – science – economy. University education is positioned at the pick of this triangle in Figure 3 because it provides a conducive environment for innovation through information and communication production and dissemination of knowledge. When this knowledge, skills, and experiences are properly acquired and utilised, NIC increases with an urge to engage in further research and development. Furthermore, there will be an increase in the desire to find a solution to human problems, leading to new ideas (innovation). All these can be adequately achieved when the tertiary institutions are efficient enough to produce a conducive environment where these knowledge, skills, and experiences are formed. This is why tertiary institutions are considered important institutional actors in a country or firm’s innovation system (Sánchez et al., 2007). In agreement with this, the tertiary institution is considered exceptional in its role in a ‘knowledge-based economy’ that creates knowledge, transmission, and dissemination (European Commission, 2005).

Figure 3. The framework of tertiary education and innovation generations and development.

Since tertiary institutions play a crucial role in developing innovative capacity, paying huge attention to these institutions becomes very important to ensure a high-efficiency level. This efficiency, however, matters most in the areas of knowledge creation (i.e., through providing experienced instructors, use of new technologies etc., and any other factor that may pose as a source of knowledge) and strengthening the R&D department. Moreover, a good management system will also go a long way in improving tertiary efficiency. Therefore, ‘investing more and better in the modernisation and quality of universities is a direct investment in the future’ (European Commission, 2005). We extend these theoretical models and contribute to the literature by (1) empirically linking national intellectual capital (NIC), tertiary education (Ter.Ed) and technological innovation (Tech.Inno) (measured by patent and trademark) in a panel of 79 nation-states from 1995–2017. (2) We also examine the channel through which efficiency in Ter.Ed affects Tech.Inno – that is, higher education has a solid ability to boost innovation activities through NIC. (3) We also contribute to innovation literature by examining additional determinants of technological progress different from existing studies. (4) The panel allows us to obtain empirical results that are particularly important in guiding policymakers across countries. (IV). We also show that the main findings are unaffected when exploring the IV quantile regression method. Finally, we also consider the sample distribution of patent and trademark applications.

The rest of this paper is designed as follows. Section 2 introduces the analysis and proposes a working hypothesis based on relevant works of literature. Section 3 describes data sources, variable selection, and descriptive statistics. Section 4 presents the empirical findings and introduces a series of robustness tests. Section 5 illustrates the possible mechanisms, and the last section concludes the paper and proposes policy implications.

2. Literature review and hypotheses development

The constructivist theory (Vanderstraeten & Biesta, 1998) enables students to build their worldview, philosophy of life, technical competence, and knowledge structures; it emphasises self-directed learning and social and situational learning experiences. This theory is essential to the idea of learning discovery. Through doing so, the students learn. The well-known educational psychologist thought that training was the cornerstone of learning and the students would be lost if they didn’t learn. He emphasised that students should build their knowledge via practice and that the psychological growth of their students should be taken into account. Gardner (1993) proposed the ‘multiple intelligence theory,’ which immediately piqued the interest of educators and spread throughout the world. This theory’s research has encouraged and directed the efficient use of educational technology to optimise customised instruction. Therefore, students use this technology as a vehicle and significant material to uncover topics, raise questions, and eventually solve them. The Theory of Invention Diffusion describes the model of receiving any innovation (Aizstrauta et al., 2015). This theory is largely used in education technology training. The spread of innovation outlines the phases of Tech.Inno, new processes and features, and recipients’ part in Tech.Inno. Understanding learners’ ability to receive new technologies can help training institutions and trainers develop and deliver training more successfully.

For a country to thrive properly in its innovative capacity, relevant institutions that enhance this innovation should be established and maintained properly (Akman & Yilmaz, 2019; Andrijauskiene & Dumčiuvienė, 2019). Some also noted theoretically that tertiary institutions are one of the most effective drives to innovation. Avvisati et al. (2014) researched educating Ter.Ed students for innovative economies. They studied the innovative capacity of two international tertiary institutions five years after graduation by accessing their respective institutions’ tertiary efficiency. This study showed that students with practical-based learning (like engineering students) are more innovative than theoretical-based students. Toivanen and Väänänen (2016) enquired whether degree acquisition directly affects the registration of patents with engineering students as a case study. They also investigated high-tech and non-high-tech startups and concluded that students from technical education would be more innovative because they have more knowledge through practical experiences. Muresan and Gogu (2012) pointed out from the statistical analysis that the role of Ter.Ed in R&D that there is a direct impact of the university’s contributions to sustainable innovation through knowledge creation. Therefore, higher education positively impacts the country’s innovation and increases its ability to embrace technological frontiers (Vandenbussche et al., 2006). Levy and Murnane (2007) explained that technological change increases the demand for expert thinking and complex communication acquired through Ter.Ed. Hanushek and Woessmann (2012) added that the value of Ter.Ed lies in its ability to provide quality education rather than mere quantitative expansion. Universities are also recognised as important institutions in economic growth and development (Ribeiro et al., 2017) because of their role in the ‘production, transmission and diffusion of knowledge,’ which builds innovative capacity (Sánchez & Elena, 2006). On this note (Todericiu & Şerban, 2015) universities are considered a ‘pipeline for NIC,’ where knowledge regarding NIC is formed and developed for innovation.

Research also proves that for a country to achieve and sustain a good innovative capacity that would spur growth and development, it must possess a buoyant NIC because it forms the knowledge, skills, and experiences that lead to innovation. Altındağ et al. (2019) carried out a literature study on the impact of IC on innovation and examined whether IC affects innovation and the existence of its mediating role. The result obtained from this study shows that a positive relationship exists between IC and innovation. Therefore, increased Ter.Ed is the resultant effect of the effective use of NIC. Chahal and Bakshi (2014) observed in their work on the ‘relationship between IC and innovation’ that IC boosts innovative organisational capacity through the employees’ skills, knowledge, and experiences. Ranking a company’s success based on its tangible and intangible materials, at all levels, the intangible assets, which consist of the IC, play more important roles than the material ones (Lin & Edvinsson, 2011). The elements of IC (human, social, and organisational capital) (Altındağ et al., 2019) create an innovative environment and equip an organisation to become proactive when transformed into information and data. Todericiu and Şerban (2015) described it as combining the stakeholder value and activities to create value in an organisation. Marcin (2013) supported this, saying that NIC is the key success factor for creating new value and determining a nation’s development. NIC is a driver for economic growth, concluding that the level of national economic development depends on the impact of IC on economic development (Díaz Díaz, 2014). NIC drives economic development radically in advanced economies due to rapid changes in new technologies they adopt, while developing countries are not paramount. Therefore, IC is considered the most important value and organisational success (Lu et al., 2014).

H0 1: National Intellectual capital has a significant positive impact on technology innovation.

H0 2: Tertiary education significantly impacts technology innovation through the channel of national intellectual capital.

3. Econometric methodology

3.1. Instrumental variable (2SLS) and instrumental variable fixed effect regression

The use of the IV or IV-FE model has two main reasons. First, it accounts for the country’s unobservable characteristics; second, it addresses the reverse causality/simultaneity problem that may occur between the dependent and independent variables. Through the instrumented process, the issue of reverse causality is solved. In addition, the IV revises for endogeneity, eliminating the inconsistencies of OLS regression (Asongu, 2012; Asongu & Mohamed, 2013). Regressor lags may be used as instruments in the first regression in the instrumental approach. (1) $$Q_{i,t}=\mathrm{ }{\alpha }_0+\mathrm{ }{\alpha }_1{Q}_{i,t-1}+\mathrm{ }{u}_{i,t}$$(1)

The instrumentation technique entails regressing the education variable on its initial lag and recording the fitted values, which are then utilised as the primary independent variable. The regressor instruments in the OLS or FE model are constructed by retaining the fitted values from OLS regression in Equation (1)(1) $$Q_{i,t}=\mathrm{ }{\alpha }_0+\mathrm{ }{\alpha }_1{Q}_{i,t-1}+\mathrm{ }{u}_{i,t}$$(1) . Heteroscedasticity and Autocorrelation Consistent (HAC) in standard errors are the requirements. The endogeneity dilemma emerges from the interaction between innovation and education since technological innovation requires students to tackle problems at a higher level of thinking. This design thinking technique identifies issues, collects data, generates potential solutions, refines ideas, and tests solutions. Innovation alters how academic institutions communicate, store, access, and utilise information, allowing them to stimulate collaboration across disciplines. The bulk of tertiary institutions’ teaching techniques has been considerably transformed by innovation. It leads to the development of novel, precise, and efficient research methodologies. Simultaneously, tertiary education plays an essential role in research and innovation as advocates of the research and innovation culture and contributors to discovering answers to diverse difficulties. Strategically using tertiary education information, abilities, and experiences results in research and innovation. Q_i,t is a ‘regressor education in country i at time t.’ α₀ is the ‘intercept,’ Q_i,t−1 is the ‘first lag of the regressor education,’ and υ_i,t is an ‘error term.’ Following that, the fitted values are employed as the explanatory variable instrument. In other words, the instrumentation approach is designed to account for non-constant variations in error terms and the likelihood that the error terms are auto-linked. This instrumentation method is consistent with previous research (Asongu & Biekpe, 2018; Efobi et al., 2019). Equation (2)(2i) $${\mathit{Patent}/\mathit{Trademark}}_{i,t}={\beta }_0+{\beta }_1{\mathit{Capital}}_{i,t}+{\beta }_2{\forall }_{i,t}+u_i+\mathrm{ }{\upsilon }_{i,t}$$(2i) grants the models of estimation according to the hypothesis. (2i) $${\mathit{Patent}/\mathit{Trademark}}_{i,t}={\beta }_0+{\beta }_1{\mathit{Capital}}_{i,t}+{\beta }_2{\forall }_{i,t}+u_i+\mathrm{ }{\upsilon }_{i,t}$$(2i) (2ii) $${\mathit{Patent}/\mathit{Trademark}}_{i,t}={\beta }_0+{\beta }_1{\mathit{Education}}_{i,t}+{\beta }_1{\mathit{education}.\mathit{capital}}_{i,t}+{\beta }_2{\forall }_{i,t}+u_i+{\upsilon }_{i,t}$$(2ii) where Equation (2i)(2i) $${\mathit{Patent}/\mathit{Trademark}}_{i,t}={\beta }_0+{\beta }_1{\mathit{Capital}}_{i,t}+{\beta }_2{\forall }_{i,t}+u_i+\mathrm{ }{\upsilon }_{i,t}$$(2i) indicates the effects of national intellectual capital on patents/trademarks. Equation (2ii)(2ii) $${\mathit{Patent}/\mathit{Trademark}}_{i,t}={\beta }_0+{\beta }_1{\mathit{Education}}_{i,t}+{\beta }_1{\mathit{education}.\mathit{capital}}_{i,t}+{\beta }_2{\forall }_{i,t}+u_i+{\upsilon }_{i,t}$$(2ii) indicates the effects of tertiary education on patents/trademarks and national intellectual capital channels. Patent/trademark represents a patent and trademark applications measuring technological innovation. Technological innovation personifies resource input and effectiveness, allowing patents and trademarks to describe intermediate outputs to increase innovation’s progressive nature (Hsu et al., 2014; Jalles, 2010). Education is the instrumented tertiary education. $\forall$ comprises the control variables: finance, income, capital, labour, population, and efficiency. α_i is a country-specific effect, and υ_i,t is the error term. (3) $${\mathit{Patent}/\mathit{Trademark}}_{i,t}={\beta }_0+{\beta }_1{\mathit{education}}_{i,t}+{\beta }_2{\mathit{finance}}_{i,t}+{\beta }_3{\mathit{income}}_{i,t}+{\beta }_4{\mathit{capital}}_{i,t}+{\beta }_5{\mathit{labour}}_{i,t}+{\beta }_6{\mathit{population}}_{i,t}+{\beta }_7{\mathit{efficiency}}_{i,t}+{\beta }_1{\mathit{education}.ca\mathrm{p}\mathit{ital}}_{i,t}+u_i+{\upsilon }_{i,t}$$(3)

Because they are resistant to ‘serial correlation, group-wise heteroskedasticity, and cross-sectional dependence,’ the Driscoll and Kraay (1998) standard errors are employed in this study. Shared shocks, regional impacts, or connections within social networks are examples of cross-sectional dependence (Baltagi et al., 2016). Estimates are skewed when cross-sectional dependence (CD) exists. Compared to other alternative covariance estimators, the Driscoll and Kraay FE model possesses small, durable sample characteristics in the presence of CD (Hoechle, 2007). As a result, it’s critical to look into the cross-sectional dependence (CD) problem in all cross-sections, whether it exists or not. Furthermore, the CD is crucial in a panel data analysis since neglecting to do so might lead to confusing and biased results. Therefore, we engaged the CD test to deal with problems such as the Pesaran cross-sectional dependence test projected by Pesaran (2015). (4) $$\mathrm{Cross}-\text{Sectional Dependence }\left(\mathrm{CD}\right)=\sqrt{\frac{2T}{N\left(N-1\right)}}\mathrm{ }\sum _{i=1}^{N-1}\sum _{j=i=1}^N{\rho }_{ij}$$(4) where T signifies time, N is the size of the panel data, and ρ_ij is the coefficient of correlation. The cross-sectional dependency test’s ‘null hypothesis is that there is no CD reliance among the cross-sectional units.’ The ‘alternative hypothesis is that CD dependency exists among the sample nations.’

3.2. Instrumental variable (IV) panel quantile regression

The research went extra by employing IV panel quantile regression. This approach considers the present degree of technical advancement in the modeling exercise. The normality assumption is not required in a PQR assessment, and evaluated parameters are generated from many positions of the dependent variable’s provisional category (Asongu, 2014a; Asongu, 2014b; Koenker & Bassett, 1978). The PQR minimises ‘the weighted sum of absolute deviation’ compared to an OLS. The θth quantile estimation of Tech.Inno is gained by resolving the given optimisation problem. (5) $$\begin{array}{c}min\\ \beta ϵR^k\end{array}\left[\begin{array}{c}\sum \theta \left|y_i-x_i\beta \right|\\ iϵ\left\{i:y_i\ge x_i\beta \right\}\end{array}+\begin{array}{c}\sum \left(1-\theta \right)\left|y_i-x_i\beta \right|\\ iϵ\left\{i:y_i<x_i\beta \right\}\end{array}\right]\mathrm{ }\mathit{where}\mathrm{ }\theta ϵ\left(0,1\right),$$(5) whereas the OLS depends on lessening the RSS, the QR minimises absolute deviations from the weighted sum. The conditional quantile of technological innovation (patent and trademark) given the regressors is: (6) $$Q_y+(\theta /X_i)=X_i{\beta }_{\theta },$$(6)

For the respective θth quantile, a unique parameter slope is modelled, and is comparable to the OLS anywhere $E(Y/X)=X_i^{\prime }\beta$ with the parameters reachable at the mean. The explained variable is Tech.Inno ($Y_i$), while $X_i$ comprises an intercept and the explanatory variables in the model. Individual slope parameters are developed (Asongu & Odhiambo, 2020). Because of the policy implications of conditional connections, quantile regression was chosen. Mean-oriented estimation is well known for providing broad inference for policies that may or may not have been effective in the past unless they are conditional based on current values of the dependent variable and fit extensively across nations with high, middle, and low levels of the dependent variable. This implies that, depending on the amount of innovation in each country, the influence of tertiary education on innovation may vary. The quantile estimation strategy’s instrumentation technique entails regressing the regressor of interest on their initial lags and then preserving the fitted values, which are subsequently used as the main regressors, as explained in the instrumental variable fixed effect section. This instrumentation method aligns with previous research (Asongu & Nwachukwu, 2017).

3.3. Data

3.3.1. Dependent variables

The first indicator of technology innovation is the ‘number of patent applications’ (Patents) filed by citizens of nations each year that WIPO finally grants. It is a privilege protected by national law and granted to the applicant within a certain time frame. We also use ‘trademark applications’ as a measure of innovation, as highlighted by scholars (Sandner & Block, 2011; Zhou et al., 2016). Most substitute goods and services can be differentiated by trademarks protected by an intellectual property right. Therefore, as an indispensable part of intellectual property rights, it has a significant advantage in measuring innovation. (Bonazzi & Zilber, 2014).

3.3.2. Explanatory variables

Higher education (tertiary education) helps more people study and advance emerging technology (Bianchi & Giorcelli, 2020). Also, information dissemination and availability are due to knowledge aggregation (Donou-Adonsou, 2019) and are expected to facilitate innovation development. Therefore, investment in academic research and other professional staff in science and preparation is vital in evaluating scientific study’s contribution to scientific advancement and innovation and implementing emerging technology and creative methods. ‘Mobile cellular subscriptions per 100 people’ indicate national intellectual capital (Lin & Edvinsson, 2013; Stevanović et al., 2018). It is projected to influence Tech.Inno positively as it signifies supporting infrastructures smoothing the access and diffusion of information. National intellectual capital, also known as process capital, comprises a nation’s non-human sources of knowledge. These sources, integrated into a country’s infrastructure, assist the generation, accessibility, and transmission of information. This form of capital is assessed by corporate competitiveness, government efficiency, intellectual property rights protection, the number of computers per capita, and the ease with which new businesses may be started. We also include other essential control variables perceived to be innovation determinants. First, the financial development index (Finance) literature suggests that the most vital role of financial development on innovation is to mitigate moral hazards and adverse selection problems (Hall & Lerner, 2010). Consequently, we control this variable using the new financial development index, which comprises financial institutions and financial markets obtained from the International Monetary Fund (IMF). Second, real GDP (Income) typically represents the overall degree of economic growth. Therefore, technical progress requires a well-anticipated socio-economic environment of overall economic development (Luo & Cheng, 2013).

Third, the population comprises the total number of people in a country. So, more people in a country can translate into more ideas, knowledge, and inventions, thus improving technological changes (Kremer, 1993). Thus, we expect a positive relationship between population and innovation capacity. We accentuate that the government’s effectiveness can stimulate innovation production (Wen et al., 2021). Moreso, the ‘Labour force participation rate;’ of the workforce is resolute by stating the number of workers as a proportion of employed persons. The higher workforce is compelled by an ever-growing intensity of competition, shifting many firms’ and nations’ strategies towards 'diversified quality production’. Therefore, the rising rate of innovation across all economies necessitates that the workforce is equipped with technical know-how and what is referred to as ‘generic skills’ – problem-solving, creativity, collaboration, and communication abilities.

4. Empirical results

4.1. Descriptive statistics

Table 1 provides the descriptive statistics of all variables. From Table 1, the average patent is 9.190; however, the median value is 8.732, with more than 50% of the countries’ innovations measured by the patent being higher than the average value. Moreover, the mean and median of the trademark are 10.154 and 9.893, respectively. We also observe the mean and median of tertiary education as 4.019 and 4.089, implying a lower difference among the countries. Furthermore, the correlation coefficients in Table A1 indicate that average tertiary education positively correlates with patent and trademark applications, respectively.

Table 1. Variables, sources, and descriptive statistics.

Variables	Definition and sources	Mean	S. D	Median
Patents	‘The total number of patents applied by country residents.’ Sources: WIPO Statistics database (2017).	9.19	1.878	8.732
Trademarks	‘The total number of trademarks applied by country residents.’ Sources: WIPO Statistics database (2017).	10.154	1.304	9.893
Education	‘Tertiary school enrollment (% gross enrollment).’ Sources: WDI (2017).	4.019	0.468	4.089
Capital	‘Mobile cellular subscriptions per 100 people’: Sources: WDI (2017).	0.713	0.128	0.728
Finance	‘Index of financial development.’ IMF (2017)	13.546	1.301	12.921
Population	‘The total number of populations residing in a particular country.’	4.163	0.96	4.521
Income	‘Real GDP at constant 2011 national prices (in mil. 2011 US$).’	3.885	0.095	3.888
Efficiency	‘Government effectiveness (ranges from 0 to 100).’ Source: WGI 2017	3.06	1.634	2.174
Labour	‘Labour force participation rate (% of the total population).’ WDI (2017).	4.504	0.138	4.541

The study utilised data from WIPO – ‘World Intellectual Property Organisation’ (WIPO Report, 2020), WDI – ‘World Development Indicator’ (World Bank, 2020), IMF – ‘International Monetary Fund’ (IMF, 2020), WGI – ‘World Governance Indicator’ (World Bank, 2020), PWT – ‘Penn World Table’ (GGDC, 2018). The data window is from 1995 to 2017 because of data disposal for the core variables.

Sources: Authors conceptualization.

5. Results and discussions

5.1. Primary findings

This section presents the conclusions of the cross-sectional dependency, 2SLS, and IV ‘fixed-effects models.’ The test for CD of the variables in the model is discussed first, which explains the appropriateness of using the ‘Driscoll and Kraay standard errors with a null hypothesis of weak CD’ of the variables and an alternate hypothesis of CD using the Pesaran (2015) procedure. The findings are shown in Table 2, which shows that the errors of the variables in the model are ‘cross-sectionally dependent’ at a 1% significance level. This explains why the standard errors of ‘Driscoll and Kraay’ were utilised in the modelling exercise. The 2SLS outcomes on the effect of NIC on Tech.Inno indicates a strong positive relationship between NIC on innovation, as seen in columns 1, 2, and 3 of Table 3. A similar finding is applicable in Table 4 on the impact of Ter.Ed on Tech.Inno. A 1% increase in Ter.Ed increases patent by 0.655 (column 1) and 1.211 (column 2) percent points, significance at 5% and 1% level, respectively, and trademark by 0.526 (column 3) percent points at 1% significance level. The IV fixed effect results in Table 5 reveal that Ter.Ed positively and significantly impacts Tech.Inno after accounting for unobserved heterogeneity. A 1% increase in Ter.Ed increases patent by 0.732 and 0.267% points, significance at 1% and 10%, respectively, and trademark by 0.205 and 0.276% points at 5% and 1% significance levels, respectively. This suggests that Ter.Ed is important for research and development and crucial for national innovations (Zambon & Monciardini, 2015). This study’s major result is that there is substantial evidence for a causal link between the availability of higher education, training, and skills and the rise in both the demand for and the availability of technical innovation. At the most basic level, it has been demonstrated that investing in capital equipment, innovation, and higher education is mostly complimentary and reinforces one another (Lloyd-Ellis & Roberts, 2002).

Table 2. Pesaran (2015) test for cross-sectional dependence.

Variables	Test values
Patent	5.489***
	(0.000)
Trademark	2.975***
	(0.003)
Education	24.586***
	(0.000)
Finance	26.756***
	(0.000)
Income	44.221***
	(0.000)
Capital	44.653***
	(0.000)
Labour	17.992***
	(0.000)
Population	36.269***
	(0.000)
Efficiency	2.845***
	(0.000)

Note: H0: Errors are weakly cross-sectional dependent. *** represents statistical significance at 1%.

Sources: Authors conceptualization.

Table 3. Instrumental variable regression for the influence of national intellectual capital on technological innovation.

Variables	Patents		Trademarks
	(1)	(2)	(3)	(4)
Capital	0.808***	0.721***	0.047	0.258***
	(0.067)	(0.107)	(0.048)	(0.082)
Finance	−2.167***	−2.578***	−0.852***	0.296
	(0.200)	(0.549)	(0.294)	(0.478)
Income	1.479***	0.611	0.965***	0.718***
	(0.027)	(0.431)	(0.046)	(0.098)
Labour	1.295***	1.958***	1.803***	1.817***
	(0.381)	(0.501)	(0.326)	(0.324)
Population		0.893**		0.187**
		(0.440)		(0.091)
Efficiency		3.870**		1.280*
		(1.544)		(0.745)
Constant	−18.47***	−28.68***	−9.758***	−12.463***
	(2.088)	(5.007)	(1.861)	(1.861)
F – statistics	3638.75	6514.08	427.01	115.31
P value	0.000	0.000	0.000	0.000
R-square	0.8763	0.8884	0.9231	0.9343
Observation, Country	308,79	266,79	305,79	263,79

1. Driscoll-Kraay robust standard errors reported in parenthesis account for cross-sectional dependency.

2. * p < 0.1, ** p < 0.05, *** p < 0.01.

Sources: Authors conceptualization.

Table 4. 2SLS (Instrumental variable regression) for the influence of tertiary education on technological innovation and the role of national intellectual capital.

Variables	Patents		Trademarks
	(1)	(2)	(3)	(4)
Education	0.655**	1.211***	0.526***	0.023
	(0.240)	(0.322)	(0.129)	(0.249)
Finance	−2.724***	−1.376**	−0.636**	0.266
	(0.352)	(0.616)	(0.277)	(0.496)
Income	1.522***	0.258	0.969***	1.013***
	(0.048)	(0.394)	(0.036)	(0.087)
Capital	0.523***	4.430***	0.376***	4.614***
	(0.121)	(1.144)	(0.104)	(0.378)
Labour	1.752***	2.840***	1.326***	1.508***
	(0.395)	(0.564)	(0.348)	(0.284)
Population		1.251***		−0.122
		(0.385)		(0.084)
Efficiency		−0.920		0.119
		(1.487)		(0.866)
Education.Capital		4.753***		4.125***
		(1.051)		(0.344)
Constant	−21.36***	−9.483	−7.174***	−5.154***
	(2.512)	(5.075)	(1.622)	(1.132)
F–statistics	1611.15	13024.00	508.64	144.26
P value	0.000	0.000	0.000	0.000
R-square	0.8795	0.9036	0.9313	0.9478
Observations, Countries	272,79	240,79	269,79	237,79

1. Driscoll-Kraay robust standard errors reported in parenthesis account for cross-sectional dependency.

2. * p < 0.1, ** p < 0.05, *** p < 0.01.

Sources: Authors conceptualization.

Table 5. Instrumental variable fixed effect regressions for the influence of tertiary education on technological innovation and the role of national intellectual capital.

Variables	Patents		Trademarks
	(1)	(2)	(3)	(4)
Education	0.732***	0.267*	0.205**	0.276***
	(0.166)	(0.163)	(0.096)	(0.104)
Finance	2.123***	2.056***	0.985**	0.690*
	(0.586)	(0.553)	(0.403)	(0.367)
Income	0.487	0.437	0.976***	1.000**
	(0.413)	(0.479)	(0.249)	(0.409)
Capital	0.642***	2.784***	0.341***	1.738*
	(0.124)	(0.328)	(0.093)	(0.916)
Labour	−0.515	−0.935	−0.218	0.332
	(0.955)	(1.062)	(1.241)	(0.828)
Population		0.648		−0.343
		(0.563)		(0.660)
Efficiency		1.488		1.503**
		(1.521)		(0.646)
Education.Capital		3.312***		1.403*
		(0.342)		(0.859)
R-squared Within	0.7517	0.7327	0.5407	0.5224
P value	0.000	0.000	0.000	0.000
F-statistics	7292.53	42941.97	3671.25	3325.66
Observations, Countries	272,79	240,79	269,79	237,79

1. Driscoll-Kraay robust standard errors reported in parenthesis account for cross-sectional dependency.

2. * p < 0.1, ** p < 0.05, *** p < 0.01.

Sources: Authors conceptualization.

We then examine the impact of the model’s other control variables. For example, financial development (Finance) has a negative and substantial coefficient in Tables 3 and 4, but when the IV fixed effect model was applied in Table 5, financial development became positive. We also see that GDP (Income) positively affects innovation when all other variables are constant in Tables 3–5. The additional finding also gives policymakers a vital implication, suggesting that the labour force is an important channel that improves national innovation by generating more ideas. We discovered a considerable positive influence of population on patents and trademarks in Tables 3 and 4, indicating that population is a vital motor for pushing Tech.Inno as creative potential develops with population size. Government efficiency (Efficiency) in patent and trademark Tables 3 and 5 are positive and significant.

5.2. Robustness test

5.2.1. Instrumental variable quantile regression

Koenker and Bassett (1978) suggest that only quantile regression can address such concerns by describing the model’s full distribution of the given dependent variable (for example, trademark or patent). We then present empirical results using specifications from Equation (5). Firstly, Figures A1 and A2 demonstrate the cumulative distribution function of innovation variables measured by patent and trademark, which exhibits a high skew. Table 6 illustrates the different quantiles of Ter. Ed’s effect on innovation. Column 1 indicates that in the 10^th quantile, education has a positive but insignificant effect on innovation performance; however, in the later stages of Ter.Ed, as displayed in columns (2)–(5), shows a one percent increase in Ter.Ed is associated with a significant increase in patents and trademarks. The signs of the estimated coefficients are consistent with the basic results. Empirical evidence from these estimations shows the positive effect of Ter.Ed on innovation is robust despite its distribution. Findings highlight the need for investment in Ter.Ed for its contributions to global Tech.Inno.

Table 6. Robustness–Instrumental variable quantile regression for the impact of tertiary education on technological innovation (Patents).

Variable	Q (0.10)	Q (0.25)	Q (0.50)	Q (0.75)	Q (0.90)
	(1)	(2)	(3)	(4)	(5)
Education	0.320	0.402*	0.532**	0.612**	0.665***
	(0.607)	(0.231)	(0.254)	(0.299)	(0.286)
Finance	2.018	1.972**	1.898***	1.853***	1.823**
	(1.369)	(0.973)	(0.571)	(0.675)	(0.870)
Income	0.261	0.427	0.688*	0.850*	0.956
	(0.959)	(0.682)	(0.404)	(0.473)	(0.610)
Capital	0.700**	0.650***	0.571***	0.521***	0.489**
	(0.343)	(0.244)	(0.144)	(0.169)	(0.218)
Labour	0.356	0.0127	−0.560	−0.907	−1.136
	(2.210)	(1.570)	(0.927)	(1.089)	(1.405)
Population	0.704	0.311	−0.306	−0.689	−0.942
	(2.244)	(1.595)	(0.945)	(1.106)	(1.427)
Efficiency	2.666	2.083	1.166	0.597	0.221
	(2.221)	(1.576)	(0.941)	(1.094)	(1.415)
Observation, Countries	240,79	240,79	240,79	240,79	240,79

1. Driscoll-Kraay robust standard errors reported in parenthesis account for cross-sectional dependency.

2. * p < 0.1, ** p < 0.05, *** p < 0.01.

Sources: Authors conceptualization.

5.3. Channel of mechanism

Based on the findings, we explore the possible channel through which Ter.Ed may affect Tech.Inno performance. The theoretical assertions in our earlier discussions state a mechanism by which Ter.Ed may affect Tech.Inno. However, we verify the existence of these mechanisms empirically as follows. We hypothesised in section 2 that there is a possible channel through which efficiency in tertiary education can affect Tech.Inno, indicates that NIC and level of education have a cause-and-effect relationship and cannot be separated from each other. In this section, we verify the existence of national intellectual capital. The relevance of NIC to productive industries, public institutions, and the general economy has drawn the attention of researchers, businesses, and organisations. The organisation’s innovation capability depends solely on its NIC or ability to utilise its knowledge resources (Kalkan et al., 2014). Productive industries have embraced more knowledge and intensive use of technology growth strategies. Still, they have also learned to incorporate NIC elements into their strategic plans to increase organisational capabilities to spur innovation. Tables 4 and 5 show that the coefficient of the interaction effect is statistically significant, indicating that NIC influences the positive impact of Ter.Ed on innovation performance. However, countries with a high degree of NIC may have a stronger ability to withstand external shocks, so Ter.Ed may have a strong and significant impact on their level of Tech.Inno. At the same time, countries with a low degree of NIC can experience a substantial decline in their level of innovation. The above results suggest that the effect of Ter.Ed might work through the channel of NIC.

‘Intellectual capital is not only an intangible asset but an ideological process’ (i.e., movement from having knowledge and skill as an asset to using them to bring forth innovation) (Chang & Hsieh, 2011). Therefore, the enterprise converts human, social, and organisational capital, NIC components, into information and data open to Tech.Inno and can act proactively. Therefore, NIC is the most essential asset to any company that wishes to create and maintain a sustainable competitive advantage in the market. However, as innovation increasingly fuels economic growth, Ter.Ed institutions face the challenges of equipping students with the skills required for innovation. Ter.Ed can provide the knowledge, skills, experiences, and other variables that can form NIC bases. Likewise, education is incomplete without NIC as an end or a means to an end. An ‘end’ when it is one of the goals of education – to increase the NIC of a country by making information available in the form of skills, knowledge, ideas etc. but a ‘means to an end’ when the components of NIC – human capital, structural capital, and relationship capital is applied in the structure of the educational system to enable it to achieve other goals etc. From the preceding, NIC is intertwined with education because universities’ main goals (Sánchez et al., 2007) are to create and disseminate knowledge. Their most important investments are in R&D and human resources, as indicated in Figure 4. Therefore, NIC reinforces the impact of Ter.Ed and improved R&D, promoting planning and development and enhancing Tech.Inno.

Figure 4. Framework on how national intellectual capital reinforces the impact of tertiary education on innovation generations.

6. Conclusion and policy implication

The Ter.Ed, research, and innovation activities are the heart of the NIC knowledge economy. However, the current state of macroeconomic literature supports a strong link between Ter.Ed and innovation mainly through theoretical arguments. This study empirically examines how technological progress responds to efficiency in Ter.Ed. Using panel data for the sample of 79 countries from 1995–2017, we show that Ter.Ed exerts a positive and strong effect on Tech.Inno measured after controlling for other determinants of Tech.Inno. Furthermore, we obtained similar estimations using an IV panel quantile regression. In particular, the empirical assessment using the 2SLS, IV fixed-effect method, and panel quantile regressions provide similar effects of Ter.Ed on Tech.Inno. In addition, we observe the impact of Ter.Ed on Tech.Inno is endowed with abundant NIC (creating new ideas) channeled into research and development, boosting human capital, opening up international trade and investment, and leading to technological progress. Theoretically, we suggest that Ter.Ed increases innovative activity by increasing NIC in R&D, promoting total investment, trade openness, and human capital development.

Ter.Ed largely influences economic development and growth through NIC and research, consistent with endogenous growth theory. As endogenous growth theory highlights, they improve information exchange, which is critical to the spread of ideas. The findings support R&D data policies that support specific digital innovation, skills, knowledge creation, and diffusion. Proactive policy frameworks should be adopted to promote national intellectual capital through internet infrastructure for economic inclusion that fosters innovativeness. Therefore, Ter.Ed, with ‘digital technologies,’ should redefine Tech.Inno, fosters partnerships and helps to form ‘innovative ecosystems.’ Promoting government and corporate funds to university researchers allows them to prioritise studies likely to have the most long-term advantages. Grants to academics may result in more patents being filed by commercial companies.

Competency-based education can advance further by leveraging technology to tailor a student’s navigation of a to-be-mastered subject. Nations should support initiatives that increase the number of persons studying ‘science, technology, engineering, and mathematics (STEM)’ through the use of ICT, Ter.Ed will help research in tertiary institutions decouple information from a ‘physical repository,’ allowing for the flow of essential information, ideas, and knowledge during development. The ease of exchanging ideas can help bridge the knowledge gap between developed and developing countries, allowing poorer countries to improve their living standards and embrace knowledge and technology transfer through FDI inflows. Therefore, nations, firms, and organisations need NIC to be equipped and foster growth and development. Nations should invest more in human capital from Ter.Ed research and development. Developing countries should increase credibility in investment in Ter.Ed to boost innovation. This will attract more investments and resource inflow that will likely spillover effects on technological progress and economic development.

Better technical change rates result in higher ‘uncertainty’ for businesses, which calls for increased flexibility and more widely dispersed problem-solving abilities. Consequently, the intensifying rivalry drives the need for a larger workforce, which can cause many businesses and governments to change their strategy in favour of ‘diversified quality production.’ Better technical change rates result in higher ‘uncertainty’ for businesses, which calls for increased flexibility and more widely dispersed problem-solving abilities. Consequently, the intensifying rivalry drives the need for a larger labour force, which will cause many businesses and governments to change their strategy in favour of ‘diversified quality production.’

Authors’ contributions

Chukwuemeka Valentine OKOLO: conceptualisation, data curation, software, formal analysis, investigation, methodology, writing–original draft. Jun WEN: visualisation, investigation, supervision, funding acquisition. Juliet Oluchi EZE: project administration, visualisation, writing–review & editing, resources.

Disclosure statement

The authors declare that they have no known competing financial interests or personal relationships that could have influenced the work reported in this paper.

Additional information

Funding

Jun WEN is grateful to the National Natural Science Foundation of China for the Project: Does Tertiary Education Promote Technological Innovation Sustainability? The Role of National Intellectual Capital. An Empirical Evidence: (the grant number: 72074176).

Appendix

Figure A1. Quantiles of patent applications. Source: Author’s computations.

Figure A2. Quantiles of trademark applications. Source: Author’s computations.

Table A1. Summary statistics.

	N	Std. Dev.	Mean	Median	min	max	var	skew	kurto	p25	p75
Patent	322	1.878	9.19	8.732	5.595	14.078	3.526	0.658	2.575	7.685	10.47
Trademark	319	1.304	10.154	9.893	8.126	14.557	1.701	0.752	3.384	9.111	11.164
Education	286	0.468	4.019	4.089	1.479	4.553	0.219	−3.011	14.377	3.925	4.288
Finance	308	0.128	0.713	0.728	0.392	0.989	0.016	−0.232	2.811	0.635	0.793
Income	322	1.301	13.546	12.921	11.543	16.759	1.693	0.501	1.995	12.44	14.684
Capital	322	0.96	4.163	4.521	−1.229	5.148	0.923	−2.27	9.301	3.958	4.774
Labour	322	0.095	3.885	3.888	3.607	4.066	0.009	−0.194	2.552	3.829	3.945
Population	322	1.634	3.06	2.174	1.295	7.251	2.669	1.056	3.341	1.701	4.175
Efficiency	266	0.138	4.504	4.541	3.765	4.605	0.019	−2.931	11.618	4.492	4.58

Sources: Authors conceptualization.

Table A2. Correlation coefficient.

	Patent	Trademark	Education	Finance	Income	Capital	Labour	Population	Efficiency
Patent	1
Trademark	0.882	1
Education	0.439	0.621	1
Finance	−0.043	−0.003	0.320	1
Income	0.898	0.945	−0.519	0.097	1
Capital	0.237	0.307	0.691	0.429	−0.237	1
Labour	0.267	0.375	−0.344	0.098	0.282	−0.017	1
Population	0.884	0.938	−0.677	−0.110	0.955	−0.368	0.261	1
Efficiency	−0.515	−0.633	0.852	0.539	−0.554	0.498	−0.248	−0.739	1

Sources: Authors conceptualization.

Table A3. List of countries.

Country	Country	Country
Albania	Greece	Pakistan
Argentina	Guatemala	Peru
Armenia	Hungary	Philippines
Australia	Iceland	Poland
Austria	India	Portugal
Azerbaijan	Indonesia	Romania
Bangladesh	Ireland	Russian Federation
Belarus	Israel	Serbia
Belgium	Italy	Slovakia Republic
Bosnia and Herzegovina	Japan	Slovenia
Brazil	Jordan	South Africa
Bulgaria	Kazakhstan	Spain
Canada	Kenya	Sri Lanka
Chile	Kyrgyzstan	Sweden
China	Latvia	Switzerland
Colombia	Lithuania	Tajikistan
Croatia	Madagascar	Thailand
Cyprus	Malaysia	Tunisia
Czech Republic	Malta	Turkey
Denmark	Mexico	Turkmenistan
Ecuador	Moldova	Ukraine
Egypt	Mongolia	United Kingdom
Estonia	Montenegro	Uruguay
Finland	Morocco	Uzbekistan
France	Mozambique	Zambia
Georgia	North Macedonia
Germany	Norway

Sources: Authors conceptualization.